Delineation of Key Regulatory Elements Identifies Points Of

DELINEATION OF KEY REGULATORY ELEMENTS IDENTIFIES POINTS OF

VULNERABILITY IN THE MITOGEN-ACTIVATED SIGNALING NETWORK

SUPPLEMENTARY MATERIALS

List of contents Supplementary Figures with legends 1. Figure S1: Distribution of primary siRNA screen data, and standardization of assay procedure. 2. Figure S2: Scatter plot of screen data. 3. Figure S3: Functional relevance of the identified targets and Calculation of residence time from PDT and cell cycle distribution. 4. Figure S4: FACS profiles for ABL1 and AKT1. Table for data in Figure 5B. 5. Figure S5: Venn diagram showing the results of the comparative analysis of other screen results 6. Figure S6: Dose response profiles for the AKT1 + ABL1 inhibitor combination for CH1, list of the 14 cell lines and their description, effect of ABL1+AKT1 inhibitor combination on increase in apoptotic cells and G1 arrest in 14 cell lines, effects of CHEK1 inhibitor on combination C1,C2 on 4 cell lines. Supplementary Tables 1. Table S1: siRNA screen results for targeted kinases and phosphatases. 2. Table S2: Gene expression status of the validated hits. 3. Table S3: Role played by identified RNAi hits in regulation of cell cycle, the effect on PDTs along with phase-specific RTs. 4. Table S4: List of molecules classified as cell cycle targets. 5. Table S5: High confidence network used for graph theory analysis. 6. Table S6: Occurrences of nodes in shortest path networks. 7. Table S7: Network file used as SNAVI background. 8. Table S8: Classification of nodes present in modules according to specificity.

Legends for tables Supplementary Experimental Procedures References

Figure S1

A 450

400 G1 S 350 G2

300

250

200

150

100

50 Distribution of molecules Distribution 0 -6-4-20246 Z-score

350 200 400 S G1 300 G2

150 300 250 200 100 200 150

100 50 100 Distribution

of molecules 50

0 0 0 -4 -2 0 2 4 -4-20246 -4-20246 Z-score

B PLK1 GAPDH

PLCg BTK

PLCg CDC2A

PLCg CHEK1

PLCg MET Distribution profiles of complete primary screen and western blots showing knockdown efficiency.

Plotted in panel A are the mean z-scores of the replicates for the entire screen with siRNA targeting kinases and phosphatases. The graph shows the distribution of the data points for all the molecules targeted. The normal nature of the distribution curve, with minimal skewness, confirms the robustness of the screen.

Panel B shows the effect of siRNA knockdown of 5 targets at 96 hours post transfection. siRNAs targeting BTK and PLK1 were used for screen standardizations. The reductions in protein levels were monitored at 24, 36, 48, 72 and 96 hours post transfection. At each of these time points cells were harvested, lysed and the levels of the respective proteins in the cytoplasm detected by Western blot analysis. Silencing of select proteins, which includes 3 validated screen hits (CHEK1, MET and CDC2A) at 96 hours after transfection is shown here. Here either GAPDH (for PLK1), or PLCg (for the remaining molecules) were also probed in parallel to serve as the loading control. The antibodies were obtained from Cell Signaling Technology. (Also refer Figure 1). Figure S2

8.00 Replicate 1 Replicate 2 G1 6.00 Chka Chka 4.00

2.00

0.00 -50 50 150 250 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 1650 1750 1850 1950 2050 2150 2250

-2.00

-4.00

-6.00

-8.00

6.00 Replicate 1 Replicate 2 S 4.00 Prkar1a Prkar1a

2.00

0.00 -50 50 150 250 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 1650 1750 1850 1950 2050 2150 2250

Z-score -2.00

-4.00

-6.00

-8.00

8.00 Replicate 1 Replicate 2

6.00 G2

4.00

2.00

0.00 -50 50 150 250 350 450 550 650 750 850 950 1050 1150 1250 1350 1450 1550 1650 1750 1850 1950 2050 2150 2250

-2.00

-4.00 MET MET -6.00

-8.00

-10.00

Molecules Scatter plot of z-scores to verify reproducibility of the primary siRNA screen Shown are the scatter plots for the two individual replicate z-scores for the three different phases. (Refer Figure 1) Figure S3

A 100 B Adenocarcinoma Gastric adenocarcinoma Lung adenocarcinoma Adenoma Adrenocortical adenoma Follicular adenoma Anaplastic astrocytoma 90 Anaplastic oligodendroma Astrocytoma Biliary tract cancer Bladder cancer Brain tumour Breast cancer 80 Breast carcinoma Cancer Lung cancer Carcinoma Ductal carcinoma Embryonal carcinoma Neuroendocrine carcinoma 70 Non small cell lung carcinoma Squomous cell carcinoma Cervical cancer/carcinoma Cholangiocarcinoma Chondrosarcoma Colon cancer Colon carcinoma 60 Colorectal cancer Colorectal carcinoma/tumor Endometrial cancer/carcinoma Epithelial tumour Esophagial carcinoma Gastric cancer/carcinoma Glioblastoma 50 Glioma Hepatocellular carcinoma

Leukemia –log (P-value) Lymphocytic leukemia T cell leukemia Lung carcinoma 40 Lymphoma B cell lymphoma Non hodgkins lymphoma T cell lymphoma Mammary tumour Melanoma Metastasis 30 Metastatic melanoma Multiple tumors Myeloid leukemia Oral cancer Osteosarcoma Ovarian cancer Ovarian carcinoma 20 Pancreatic cancer Pancreatic carcinoma Primary tumor Promyelocytic leukemia Prostrate cancer Prostrate carcinoma 10 Retinoblastoma Rhabdosarcoma Sarcoma Skin cancer Solid tumors Thyroid cancer Tumors 0

Functional relevance of the identified targets and Calculation of residence time from PDT and cell cycle distribution Panel A depicts the gene-disease relationship score (indicated by the color bar) of our hits with various forms of cancer. These scores were obtained as –log (P-value) from Novoseek gene-disease database.

Panel B shows the schematic to explain the method followed to calculate RT for CH1 cells after targeted siRNA mediated perturbation of screen hits. The approach is illustrated by taking the case of PLK1 as a typical example. (Refer Figure 1C and 1D) Figure S4 A # Cells

Panel A shows the histograms obtained from a cell cycle analysis of CH1 cells at 48 hours post treatment with 0.5 x IC50 of ABL1 and AKT1 inhibitor (i.e Imatinib mesylate and LY294002). The respective RTs of the G1, S and G2 phases are indicated. (Figure 5A). Panel B lists the siRNA sequences employed for silencing of the high stress and betweeness targets shown in Figure 5B. Figure S5

A B

C D E

Panel A shows the Venn diagram comparing the overlap of IMP nodes identified from the three different screens. Panel B shows the Venn diagram comparing the overlap of nodes from the IMP node networks of the three different screens. Panel C is the comparison between IMP nodes identified from three sets of randomly sampled source-target pairs. A very poor overlap can be observed here. Panel D is the comparison between IMP nodes identified from three sets of randomly sampled sources but retaining the target nodes to assess target biasness in IMP nodes selection. Again a very poor overlap can be observed here. Panel E is the comparison between IMP nodes identified from three sets of randomly sampled sources and the IMP nodes identified from our study in CH1. A very poor overlap can be observed here when compared to the overlap observed between U2OS and HeLa cells. Figure S6

A B

G1 population Dead cells 70 70

60 AAKT1 60 AABL1 AAKT1+ABL1 50 50 40

40 30 20

% % Apoptotic cells 30 10

20 0 0.01 0.1 1 10 0.01 0.1 1 10 Dose (x IC50)

C % % Apoptotic cells

DE G1 % cells % G1 Δ % % Apoptotic cells Panel A depicts the dose-response profile for increase in G1 population, and increase in apoptotic cells in CH1 cells treated with pharmacological inhibitors of ABL1, AKT1 and the combination of the respective doses. Doses for each inhibitor used were in multiples of their corresponding IC50 values as indicated. Values (mean ± S.D. of three experiments) are expressed as the percent of cells obtained at 48h post treatment. (Also refer Figure 5).

Panel B lists and describes all the human cell lines used in this study.

In Panel C, the effect of inhibition of ABL1, AKT1 and their combination (at concentrations corresponding to 5 times their respective IC50 values) on inducing apoptosis of fourteen different human cancer cell lines is shown. Here, the cells were treated with a single addition of the inhibitor or inhibitor combination, and the frequency of apoptotic cells was determined 48h later. In all cases, results are the mean (± S.D.) of three experiments.

In Panel D, cells of the indicated cancer cell lines were treated with inhibitors against either the indicated kinase, or the kinase combinations and the consequent effect on cellular apoptosis was determined 48 hours later. The concentration of the relevant inhibitor employed was five-fold greater than its corresponding IC50 value in all cases and results are the mean (± S.D.) of three experiments. The protocol followed was similar to panel C. As is evident, with the exception of Jurkat cells, apoptosis resulted from the dominant effects of CHEK1 inhibition in all cases (Refer Figure 7).

Panel E depicts the increase in G1 population observed in the fourteen different human cancer cell lines following treatment with inhibitors of both ABL1 and AKT1 (at concentration corresponding to their IC50 values). Values (mean ± S.D., n=3) are those obtained at 24 h post-treatment and are expressed as increase in the percent of cells in the G1 phase, relative to that in the absence of any inhibitor (i.e. vehicle only). Table S3 Classification of hits based on RT affecting the cell cycle phases

Observed Population Phenotype Total No. GO Involved Doubling (Based on of GO Cell Cycle or time increase in RT) Pubmed IDs for Cell Cycle role known Role in cell proliferation terms Proliferation Ratio Residence Time (Hours)(Hours) Std. Dev (± hours) G1 and G1/S S G2 G1 S G2 PDT G1 (SD) S (SD) G2 (SD) PDT (SD) 1 Atm G1 7923116 11544175:8843193 -- -- 13 5 0.4 10.2 11.5 3 23.9 0.5 0.7 0.2 1.5 2 Chek1 G1 11544175 11544175 11544175 -- 4 3 0.8 12.2 15.2 3.1 29.5 0.8 1.1 0.3 2.3 3 Fgfr2 G1 ------10572038:15199177 55 3 0.1 11.7 10 3.8 25.6 1.6 1.7 0.6 3.9 4 Irak2 G1 ------3 0 0.0 10.7 11 3.8 26.5 1.1 1.3 0.5 2.9 5 Lek1 G1 14555653 ------14 4 0.3 13.9 14.2 3.8 31.4 1 1.2 0.3 2.6 6 Mast2 G1 ------3 0 0.0 14.7 13.2 2.8 32.8 1.4 1.4 0.4 3.2 7 Pkig G1 ------2 0 0.0 9.6 13 4.1 25.8 0.8 1 0.4 2.2 8 Pkn1 G1 ------4 0 0.0 10.1 13.4 3.1 26.3 0.7 1.1 0.4 2.2 9 Plk2 G1 12972611 -- -- 12972611 2 1 0.5 14.6 14.3 3.9 31.8 1 1.3 0.3 2.6 10 Ripk3 G1 ------10358032 4 2 0.5 10.5 13.5 2.2 26.1 0.5 0.7 0.2 1.4 11 Tex14 G1 ------1 0 0.0 11 14.8 1.9 27.5 1.6 2.2 0.5 4.2 12 Vrk2 G1 18286197 -- -- 1 0 0.0 12.1 11.9 3.5 27.5 1.3 1.4 0.6 3.3 13 Cdk4 G1 and G2 19013283 --- 16476733:10318807 --- 4 2 0.5 8.7 12.7 4.8 26.4 0.6 0.9 0.3 1.8 14 Cdc25a G1 and S 14681206:10454565 10454565:10995786 19192479 -- 2 2 1.0 11 16.9 3.2 30.5 1 1.5 0.4 2.9 15 Cdc2a G1 and S 17700700 17635936 19570916:19013283 -- 3 3 1.0 14.9 14.9 3.4 32 0.9 1.2 0.3 2.4 16 Cdk5 G1 and S 8302605:1358458 ------6 2 0.3 11.5 19.5 3.3 33.6 0.8 1.2 0.3 2.3 17 Chka G1 and S 18296102 ------1 0 0.0 17.7 17 3.1 37.8 1.4 1.7 0.4 3.5 18 Dusp12 G1 and S ------16274472 4 0 0.0 10.3 20.2 3.9 34.4 0.7 1.3 0.3 2.4 19 Map3k5 G1 and S ------15210709 1 0 0.0 11.9 25.4 3.4 41.4 1.3 2.3 0.5 4.1 20 Map4k1 G1 and S 14981547 15634691 -- -- 3 0 0.0 9.8 14.9 2.8 28.1 0.9 1.4 0.4 2.7 21 Met G1 and S 15064724 ------12 0 0.0 14 16.4 2.6 33.5 1.7 2.1 0.6 4.4 22 Npr1 G1 and S ------3 0 0.0 9.4 14 3.1 27 0.8 1.3 0.4 2.6 23 Prkar1a G1 and S ------16569736 6 2 0.3 9.4 16.2 6.2 29.9 0.8 1.3 0.7 2.9 24 Slk G1 and S -- -- 16236704 --- 5 1 0.2 10.4 24 3.1 37.5 0.3 0.5 0.1 0.9 25 Srms G1 and S ------1 0 0.0 10.9 16.7 1.2 28.6 1.8 2.6 0.4 4.9 26 Taok2 G1 and S ------10660600 1 0 0.0 13.6 17.1 1.4 31.8 1.4 1.7 0.3 3.4 27 Trpm7 G1 and S 19515901 ------18590694:19515901 1 0 0.0 11 16.2 2.1 29.2 1.3 1.8 0.5 3.7 28 Tssk5 G1 and S ------1 0 0.0 9.8 15.2 1.8 26.6 1.4 2.1 0.4 3.9 29 Csnk2a1 G2 12396231 --- 12396231 --- 2 0 0.0 6.1 12.7 7.6 26.1 0.9 1.4 0.8 3.1 30 Prkch G2 16434969 --- 8454583:8336936: 9310352 1 0 0.0 3.4 13 12.7 28.3 0.9 1.9 1.6 4.4 31 Prkg1 G2 ------11073964 5 0 0.0 7.6 15.3 7.2 28.6 0.9 1.7 1 3.6 32 Prkg2 G2 ------1 0 0.0 8.1 15 7 28.5 0.9 1.5 0.8 3.1 33 Clk4 S ------2 0 0.0 6.8 21.5 5.2 33.1 0.9 2.3 0.7 3.9 34 Ephb2 S ------15973414 6 1 0.2 6.1 21.6 1.4 28.9 0.5 1.6 0.2 2.4 35 Guk1 S ------1 0 0.0 9.4 29.7 4.2 43.1 0.8 2 0.4 3.3 36 Itgae S ------2 0 0.0 7.4 18.3 1.9 28.2 0.7 1.3 0.3 2.3 37 Mapkap1 S ------17012250 1 0 0.0 7.5 15.9 3.4 27.1 0.5 1.1 0.3 1.9 38 Neo1 S ------3 0 0.0 7.6 17.3 3.3 28.6 0.7 1.4 0.4 2.5 39 Ppm1g S -- 17054950 -- --- 1 1 1.0 6.3 14.8 2.1 23.1 0.5 1 0.2 1.7 40 Ppm1h S ------18059182 2 0 0.0 8.6 14.5 1.9 25.5 1.2 2.2 0.4 3.8 41 Taf9 S -- 12837753 -- 12837753 5 2 0.4 8.1 15.5 2 25.6 1.1 1.7 0.3 3.1 42 Inpp5f S and G2 ------1 0 0.0 9.8 17.5 6.8 34.2 1 1.3 0.5 2.8 43 Plk1 S and G2 -- --- 15838519:16557283 --- 3 3 1.0 9.4 19.3 8.6 36.6 1.1 2 0.7 3.8 8.4 13.5 4.1 26 1.5 1 0.5 2.5 --- Role in Cell cycle/Proliferation not well known

The table provides the classified list of molecules based on their exclusivity for regulation of a particular phase of the cell cycle. Also listed are the ratio of observed versus predicted functions of the hits using GO terms. Table S4 List of molecules classified as cell cycle targets.

G1 G2 PMIDs S TARGETS PMIDs PMIDs TARGETS TARGETS 1 ATM 11544175:8843193 1 ATM 11544175:8843193 1 ANAPC2 11285280 2 ATR 11544175:18606783 2 ATR 11544175:18606783 2 ATM 11544175 3 CCNC 1833066:7568034:9552396:8730095 3 CCNA1 11566717:19167337 3 ATR 11544175 4 CCND1 19167337 4 CCNA2 11566717:19167337 4 AURKB 19342897 5 CCND2 19167337 5 CCNC 1833066:7568034:9552396:8730095 5 CCNB1 11285280 6 CCND3 19167337 6 CCNE1 19167337:7797073:8156587 6 CCNB2 9926943 7 CCNE1 19167337:7797073:8156587 7 CCNE2 19167337:7797073:8156587 7 CCNB3 7474080 8 CCNE2 19167337:7797073:8156587 8 CCNH 7533895:9130709 8 CCNG1 16322753 :18408012 9 CCNG1 12556559 9 CDC25A 10454565:10995786 9 CCNH 7533895:9130709 10 CCNG2 11956189:9139721 10 CDC2A 17635936 10 CDC16 12629511 11 CCNH 7533895:9130709 11 CDC45L 19054765 11 CDC20 11285280 12 CDC25A 14681206:10454565 12 CDC6 16439999:18687693 12 CDC25A 19192479 13 CDC2A 17700700 13 CDC7 10846177 13 CDC25C 18604163 14 CDC34 8248134:11514588 14 CDK2 19167337 14 CDC2A 19570916:19013283 15 CDC37 8703009:9228064:14701845 15 CDK7 9832506:7929589 15 CDK10 8208557 16 CDC7 10846177 16 CDK8 9584184:17612495 16 CENPA 18007590 17 CDK2 19167337 17 CHEK1 11544175 17 CENPB 16183641 18 CDK4 19167337 18 CHEK2 11544175 18 CENPC 12490152 19 CDK5 8302605:1358458 19 E2F1 10619603:19603422 19 CENPE 9763420 20 CDK6 19167337 20 E2F2 19457859 20 CENPH 16875666 21 CDK7 9832506:7929589 21 E2F3 15467444 21 CHEK1 11544175 22 CDK8 9584184:17612495 22 E2F4 19585502:19562678 22 CHEK2 11544175 23 CDKN1C 7729684 23 E2F5 15467444 23 MAD2L1 19461085 24 CDKN2B 19167337 24 E2F6 17457055:19549334 24 MAD2L2 10366450 25 CDKN2C 19167337 25 FOS 17216194 25 p27 17954563: 16310807 26 CDKN2D 19167337 26 GMNN 19487297:11125146 26 PLK1 19033445 27 CHEK1 11544175 27 JUN 17216194 27 PLK2 19001868 28 CHEK2 11544175 28 JUNB 18793152 28 PLK3 11971976 29 E2F1 10619603:19603422 29 MCM2 17680271 29 STK6 19342897 30 E2F2 19457859 30 MCM3 19064704:19377303 31 E2F3 15467444 31 MCM4 16754955 32 E2F4 19585502:19562678 32 MCM5 16754955 33 E2F5 15467444 33 MCM6 16754955 34 E2F6 17457055:19549334 34 MCM7 16754955: 19647517 35 FOS 17216194 35 MDM2 15485814:8628312:10995885:9087426 36 HIPK2 12851404 36 MNAT1 10026172:11113200 37 JUN 17216194 37 MYC 19554081 38 JUNB 18793152 38 ORC1L 19197067 39 MDM2 15485814:8628312:10995885:9087426 39 ORC2L 17680271 40 MNAT1 10026172:11113200 40 ORC3L 17716973 41 MYC 12631706 41 ORC4L 18652488:19647517 42 p27 8730099:10385618:8033213 42 ORC5L 17716973 43 PLK2 12651910:12972611 43 ORC6L 17716973 44 RB1 11018009:7777526:9694794 44 PCNA 19595719 45 RBL1 1152119 45 RB1 11018009:7777526:9694794 46 RBL2 1152119:9188854:8253384 46 RBL1 1152119 47 TFDP1 14618416:18687693 47 RBL2 1152119:9188854:8253384 48 TFDP2 9704927 48 RIS2 18006686:11125146 49 Trp53 7742528:9843965:11884608:10072388 49 SKP1A 18353424 50 Trp63 17114587:14737098 50 SKP2 19477924 51 TRP73 10562283 51 TFDP1 14618416:18687693 52 TFDP2 9704927 53 Trp53 7742528:9843965:11884608:10072388

List of molecules classified as cell cycle targets. Table S4 provides the complete details of classification of molecules according to their functional role in cell cycle and their corresponding PMID for the literature evidence. Table S6 Occurrences of nodes in shortest path networks.

G1 Phase S Phase G2 Phase G1S Phase No. of No. of No. of No. of Molecule occurances Score Molecule occurances Score Molecule occurances Score Molecule occurances Score TRP53 612 10.8 TRP53 372 8.2 LASP1 175 8.9 CDK2 833 11.4 RB1 471 8.3 ITGB7 367 8.1 PRKD1 159 8.1 TRP53 648 8.8 PTK2B 457 8.0 NTN1 278 6.1 TRP53 137 7.0 ABL1 465 6.3 PRKCA 385 6.8 ADORA2B 278 6.1 PRKCA 126 6.4 POLD1 463 6.3 TRAF6 332 5.8 FOS 229 5.0 CDK7 74 3.7 ANXA1 434 5.9 TICAM2 332 5.8 ABL1 225 4.9 FZR1 60 3.0 CREB1 424 5.8 KIF23 316 5.5 YWHAG 204 4.5 EGFR 59 2.9 PRKG1 391 5.3 YWHAQ 313 5.5 PIN1 201 4.4 POLD1 55 2.7 GRB2 388 5.3 BCL2 266 4.6 TRAF6 190 4.1 CDK2 46 2.3 EGFR 344 4.7 CDK2 258 4.5 CSE1L 189 4.1 CHEK2 41 2.0 CREBBP 311 4.2 IKBKAP 251 4.4 CDK2 177 3.9 ETS2 40 2.0 RIS2 294 4.0 MAPK8 234 4.1 DOK1 154 3.3 PRKG1 34 1.7 PRKCA 289 3.9 ESR1 195 3.4 DCC 151 3.3 AREG 34 1.7 SRC 273 3.7 POLD1 192 3.3 AREG 141 3.0 CASP3 31 1.5 RB1 268 3.6 PTEN 191 3.3 POLD1 138 3.0 NFE2L2 28 1.3 UBB 260 3.5 CREBBP 184 3.2 RIS2 138 3.0 CREBBP 27 1.3 AKT1 238 3.2 AKT1 178 3.1 STAT1 138 3.0 ACTB 26 1.2 EP300 236 3.2 EP300 169 2.9 E7 135 2.9 HDAC1 25 1.2 BRCA1 230 3.1 MCM7 160 2.8 EZR 128 2.8 CTNNB1 25 1.2 MCM7 230 3.1 SRC 153 2.6 PKM2 127 2.7 CREB1 24 1.1 ESR1 212 2.8 AREG 147 2.5 MCM7 125 2.7 UBTF 23 1.1 CTNNB1 191 2.6 EPC1 146 2.5 BRCA1 106 2.3 SYK 22 1.0 POLR2A 181 2.4 UBB 145 2.5 PRKCA 103 2.2 ZYX 21 1.0 STAT1 180 2.4 POLR2A 130 2.2 SMAD3 99 2.1 ESR1 21 1.0 MAPK3 174 2.3 ARHGEF7 124 2.1 RB1 93 2.0 NR3C1 20 0.9 CCNA2 166 2.2 ZBTB17 118 2.0 EED 82 1.7 MNAT1 19 0.9 AREG 165 2.2 BRCA1 115 2.0 TAF1 75 1.6 YWHAQ 19 0.9 LYN 162 2.2 EEF1D 115 2.0 POLR2A 75 1.6 IKBKAP 19 0.9 GADD45G 152 2.0 CASP3 114 2.0 EP300 70 1.5 GTF2I 18 0.8 SKP2 149 2.0 GRB2 103 1.8 UBB 70 1.5 DIAPH2 18 0.8 CDKN1B 148 2.0 SYNGAP1 100 1.7 RBL1 70 1.5 APC 18 0.8 CSNK2A1 137 1.8 RAF1 95 1.6 CCNA2 68 1.4 SYNGAP1 17 0.8 ELK1 136 1.8 WT1 94 1.6 RAF1 67 1.4 CDC20 17 0.8 JUN 136 1.8 CSNK2A1 94 1.6 DAB1 67 1.4 DLGAP4 17 0.8 CDC25A 136 1.8 TICAM1 92 1.6 ETF1 67 1.4 PPP1R14A 16 0.7 TOPBP1 127 1.7 NFKBIA 92 1.6 GRIN2B 66 1.4 MAPK8 16 0.7 E2F1 126 1.7 ABL1 90 1.5 COIL 66 1.4 ACTC1 15 0.7 TUBA4A 119 1.6 E2F1 89 1.5 MAPK1 65 1.4 E6 15 0.7 MYC 119 1.6 DLGAP4 89 1.5 CREBBP 60 1.2 RHOD 14 0.6 EPC1 114 1.5 MYC 89 1.5 ABL2 56 1.2 AKT1 14 0.6 CASP3 113 1.5 LCK 88 1.5 CASP3 56 1.2 MAPK1 14 0.6 WT1 105 1.4 XRCC6 86 1.5 MCM3 55 1.1 UBB 12 0.5 PTPN2 104 1.4 TANK 86 1.5 EPC1 55 1.1 BRCA1 12 0.5 CCND1 103 1.3 TRIM27 82 1.4 JUN 54 1.1 ITPR1 12 0.5 ARHGEF7 100 1.3 MAPK1 80 1.3 CDK5 51 1.0 ATR 12 0.5 YWHAE 100 1.3 PRC1 76 1.3 YWHAE 51 1.0 CSNK2A1 12 0.5 GADD45A 98 1.3 RACGAP1 76 1.3 EGFR 50 1.0 RB1 12 0.5 PCNA 96 1.3 SMAD3 76 1.3 E2F1 49 1.0 TUBA4A 11 0.5 MCM5 94 1.2 YWHAE 71 1.2 MCM5 46 0.9 CIB1 11 0.5 ZBTB17 93 1.2 LIMA1 71 1.2 ESR1 44 0.9 SHC 11 0.5 PPP2R4 90 1.2 CDKN1B 70 1.2 LYN 42 0.8 MCM7 11 0.5 PITPNM1 88 1.1 MRGBP 64 1.1 MYC 42 0.8 NPM1 11 0.5 MAPT 77 1.0 CCNA2 63 1.0 TOPBP1 40 0.8 PPP1CA 11 0.5 PRKCD 76 1.0 PCNA 62 1.0 RBL2 40 0.8 STAT3 10 0.4 MAPK8 75 1.0 MNAT1 61 1.0 UBA1 39 0.8 ABL1 10 0.4 GTF2I 74 0.9 MDM4 59 1.0 ITGA4 36 0.7 IQGAP1 9 0.4 ORC1L 72 0.9 MYOD1 58 1.0 AKT1 35 0.7 ACTA1 8 0.3 MAPK14 70 0.9 CDK5 57 0.9 PICK1 35 0.7 WAS 8 0.3 CDK5 70 0.9 ATM 57 0.9 SRC 34 0.7 XRCC6 8 0.3 TRIM27 68 0.9 RBL1 57 0.9 TRIM27 33 0.6 CCNA2 8 0.3 RBL1 67 0.9 MDM2 56 0.9 SKP2 33 0.6 CSNK1A1 8 0.3 CHKA 65 0.8 EGFR 56 0.9 VHL 33 0.6 HDAC2 8 0.3 E7 64 0.8 RBL2 55 0.9 LCK 32 0.6 SKP1A 7 0.3 KRT8 63 0.8 CREM 55 0.9 ORC1L 32 0.6 MDC1 7 0.3 MYOD1 62 0.8 GSK3B 52 0.9 PPP2CA 32 0.6 CDC25B 7 0.3 NPM1 60 0.8 PPP2R4 51 0.8 TLN1 28 0.5 MYC 7 0.3 NFKBIA 59 0.7 SMAD2 51 0.8 CTNNB1 28 0.5 MAPK3 7 0.3 SMAD3 56 0.7 CTNNB1 51 0.8 TFDP2 27 0.5 MDM4 7 0.3 STAT3 55 0.7 PAK4 49 0.8 PTK2 27 0.5 BUB1B 7 0.3 XPO1 54 0.7 FGFR1 48 0.8 TFDP1 26 0.5 TOP1 7 0.3 IKBKAP 54 0.7 SHC 47 0.8 XPO1 26 0.5 NCK1 6 0.2 MAP3K3 52 0.6 CHKA 47 0.8 NME6 25 0.5 PMVK 6 0.2 DLGAP4 52 0.6 ATR 44 0.7 ORC2L 25 0.5 PLCG2 6 0.2 PPP1CA 52 0.6 KAT5 43 0.7 CDKN1B 24 0.4 PCNA 6 0.2 SYNGAP1 51 0.6 STAT3 42 0.7 PPP1CA 24 0.4 PIK3R1 6 0.2 MCM3 49 0.6 PIAS4 41 0.7 YY1 23 0.4 HAND2 6 0.2 MAPK1 47 0.6 CCND1 40 0.6 MNAT1 23 0.4 PLCG1 6 0.2 HIST3H3 47 0.6 BIRC6 39 0.6 CDK9 23 0.4 ATM 6 0.2 MRGBP 46 0.6 JUN 39 0.6 MAPT 23 0.4 JUN 6 0.2 NR3C1 46 0.6 PLK1 38 0.6 HDAC1 22 0.4 DAXX 6 0.2 FHL2 46 0.6 MAPK14 38 0.6 MRGBP 22 0.4 SRC 6 0.2 YWHAB 46 0.6 E7 38 0.6 ACP1 21 0.4 HSP90AA1 6 0.2 MNAT1 44 0.5 LYN 37 0.6 ARHGEF7 21 0.4 PARP1 6 0.2 CDK4 43 0.5 BCL2L1 37 0.6 DNMT1 21 0.4 PRKAR2B 5 0.2 GSK3B 43 0.5 FHL2 37 0.6 XRCC6 21 0.4 STK11 5 0.2 CSNK1D 42 0.5 NR3C1 36 0.6 MAPK3 21 0.4 NEDD9 5 0.2 SHC 41 0.5 CHEK2 35 0.6 MCC 21 0.4 SMARCA4 5 0.2 PLK1 41 0.5 MAPK9 34 0.5 NCOA3 20 0.4 EIF4EBP1 5 0.2 CCNE1 41 0.5 SLC6A4 33 0.5 CDC2L1 20 0.4 THRA 5 0.2 SMARCA4 41 0.5 TFDP2 32 0.5 DNMT3A 20 0.4 LCK 4 0.1 MCM4 38 0.5 CDK4 32 0.5 PKN2 20 0.4 BLM 4 0.1 CSNK1A1 38 0.5 SNCA 32 0.5 CSNK2B 19 0.3 CDK5 4 0.1 SMAD2 37 0.4 TFDP1 30 0.5 SP1 19 0.3 TGM2 4 0.1 SP1 37 0.4 CASP8 30 0.5 TBP 19 0.3 AKAP1 4 0.1 PIK3R1 37 0.4 EPHA4 30 0.5 CCNA1 18 0.3 PLK1 4 0.1 LCK 36 0.4 UBE2I 29 0.4 CSNK1A1 18 0.3 FHOD1 4 0.1 DGKZ 36 0.4 UBC 29 0.4 RELA 18 0.3 ITGB3 4 0.1 TFDP1 35 0.4 GADD45G 28 0.4 LRRC59 18 0.3 CCND1 4 0.1 ORC2L 35 0.4 HDAC1 28 0.4 DDB1 17 0.3 GRB2 4 0.1 KIT 35 0.4 HSP90AA1 28 0.4 GNB2L1 17 0.3 BCL2 4 0.1 MAP2K1 35 0.4 DYNLL1 27 0.4 HDAC2 17 0.3 NDC80 4 0.1 KPNA2 34 0.4 DGKZ 27 0.4 EPB41 16 0.3 CDC25C 4 0.1 XRCC6 34 0.4 MED29 25 0.4 SMAD2 16 0.3 CHKA 4 0.1 ITPR1 34 0.4 PRG2 25 0.4 RAD52 16 0.3 HSP90AB1 4 0.1 HSP90AA1 34 0.4 LRRC59 25 0.4 SUV39H1 16 0.3 BUB1 4 0.1 PKN2 33 0.4 MAPT 25 0.4 MCM4 16 0.3 PPP1R9B 4 0.1 TFDP2 32 0.4 YWHAZ 25 0.4 NEK6 16 0.3 VAV1 4 0.1 BCL2 32 0.4 BLM 25 0.4 GADD45G 15 0.2 NCKIPSD 4 0.1 S100B 32 0.4 AATF 25 0.4 TAF11 15 0.2 ACTN4 3 0.1 CREM 32 0.4 CCNA1 24 0.4 MED9 15 0.2 PITPNA 3 0.1 EPB41 31 0.4 TOP2A 24 0.4 TCF7L2 14 0.2 EP300 3 0.1 UBE2I 31 0.4 HDAC2 24 0.4 GLCE 14 0.2 INSR 3 0.1 TUBB 31 0.4 TFAP2A 24 0.4 DUT 14 0.2 VASP 3 0.1 RPS6KB1 31 0.4 CRK 24 0.4 UBC 14 0.2 RAF1 3 0.1 MAP3K7 30 0.3 PDPK1 24 0.4 MED1 13 0.2 PTPN1 3 0.1 TRP73 30 0.3 GART 24 0.4 SORBS2 13 0.2 MAP3K7 3 0.1 NCOA6 30 0.3 PPP1CA 24 0.4 CDC25A 13 0.2 MAP3K3 3 0.1 MED29 28 0.3 MAP3K3 23 0.3 ITGB5 12 0.2 TELO2 3 0.1 NCK1 28 0.3 FGF8 23 0.3 CENPJ 11 0.2 PALM 3 0.1 RBL2 28 0.3 RELA 23 0.3 KIT 11 0.2 BRCA2 3 0.1 MAPK8IP3 27 0.3 HSPA8 23 0.3 NCOA2 11 0.2 MIS12 3 0.1 PRKAR1A 26 0.3 PTPN2 23 0.3 MPP1 11 0.2 KPNA3 3 0.1 ATM 26 0.3 TBP 23 0.3 GADD45A 11 0.2 HIST3H3 3 0.1 RPS6KA3 26 0.3 FLNA 23 0.3 CDC5L 11 0.2 CDKN1B 3 0.1 IGF1R 25 0.3 PML 22 0.3 PRKCB1 11 0.2 PTP4A3 3 0.1 ATR 25 0.3 MED1 22 0.3 PITPNM1 11 0.2 FRAP1 3 0.1 CDC6 25 0.3 CDC2L5 22 0.3 YWHAB 11 0.2 CCNB1IP1 3 0.1 CRK 25 0.3 CDC2L1 21 0.3 LMNA 10 0.1 KAT5 3 0.1 PTP4A3 25 0.3 SORBS2 21 0.3 PARD3 10 0.1 E2F1 3 0.1 SMAD4 25 0.3 CTBP1 21 0.3 CAPN1 10 0.1 CALM1 3 0.1 E6 25 0.3 YAP1 21 0.3 PTPN2 10 0.1 CRIP2 3 0.1 MDM4 25 0.3 TRAF1 20 0.3 HCFC1 10 0.1 CRKL 3 0.1 PIN1 24 0.3 MAP3K7 20 0.3 CCNH 10 0.1 FLNA 3 0.1 RAF1 24 0.3 PITPNM1 20 0.3 SIAH1 10 0.1 PTPN13 3 0.1 DDB1 23 0.2 CDC25B 19 0.3 MAPK14 9 0.1 HIST1H3C 3 0.1 SLC6A4 23 0.2 CCNH 19 0.3 PFAS 9 0.1 MARK3 3 0.1 RAD52 23 0.2 MED6 19 0.3 NCOR2 9 0.1 MAD2L2 3 0.1 CDC7 23 0.2 PCAF 18 0.2 ITGB3 9 0.1 POLR2A 3 0.1 MYBL2 22 0.2 GADD45A 18 0.2 TRP73 9 0.1 PITPNM1 3 0.1 CCNH 22 0.2 SP1 18 0.2 ATM 9 0.1 YWHAB 3 0.1 SNCA 22 0.2 NPM1 18 0.2 BCL2 9 0.1 HDAC3 3 0.1 KAT5 22 0.2 STK11 18 0.2 CRKL 9 0.1 YY1 2 0.0 TBP 22 0.2 NCK1 18 0.2 E6 9 0.1 IL16 2 0.0 VAV1 22 0.2 BMPR1B 18 0.2 PRKCD 9 0.1 VCL 2 0.0 TOP2A 21 0.2 E6 18 0.2 MDM4 9 0.1 NCOA3 2 0.0 CDX2 21 0.2 MDC1 17 0.2 MYOD1 9 0.1 KDR 2 0.0 MAPK9 21 0.2 EPB41 17 0.2 ROCK1 8 0.1 RAC1 2 0.0 IRS1 20 0.2 RIPK1 17 0.2 SLC33A1 8 0.1 MAFG 2 0.0 MCM6 20 0.2 CDC25C 17 0.2 CRK 8 0.1 RAD1 2 0.0 TP73L 19 0.2 FOXO1 17 0.2 MED29 8 0.1 PPP2R4 2 0.0 MED1 19 0.2 PFAS 16 0.2 ITGB2 8 0.1 CDC2L5 2 0.0 HDAC1 19 0.2 PTPN12 16 0.2 KAT5 8 0.1 CCNB1 2 0.0 GADD45B 19 0.2 CHEK1 16 0.2 CDKN2A 8 0.1 RAD51 2 0.0 GNB2L1 18 0.2 VIM 16 0.2 NF2 8 0.1 VIM 2 0.0 CSK 18 0.2 MAP2K1 16 0.2 NFATC2 8 0.1 PRKAR1A 2 0.0 CDC2L5 18 0.2 ELF3 15 0.2 NUMB 8 0.1 LYN 2 0.0 PTPN1 17 0.2 SLC9A3 15 0.2 VDR 7 0.1 GDNF 2 0.0 PML 17 0.2 GTF2I 15 0.2 PML 7 0.1 PSMA3 2 0.0 RPS6KA1 17 0.2 RUVBL1 15 0.2 USF2 7 0.1 SUMO2 2 0.0 MED6 17 0.2 VHL 15 0.2 DAB2 7 0.1 GADD45A 2 0.0 LASP1 17 0.2 FGFR3 14 0.2 SDCBP 7 0.1 BAD 2 0.0 PLCG2 16 0.2 PSMD10 14 0.2 RUVBL1 7 0.1 ASF1A 2 0.0 RXRA 16 0.2 PKN1 14 0.2 MED6 7 0.1 PCYT1A 2 0.0 CDC25C 16 0.2 RAD17 14 0.2 GSK3B 7 0.1 TRP73 2 0.0 PKN1 16 0.2 CLSPN 14 0.2 CAMK2G 7 0.1 YWHAE 2 0.0 POLA1 16 0.2 CSNK1D 14 0.2 KCNK3 6 0.0 ARRB1 2 0.0 YY1 16 0.2 TNFAIP3 14 0.2 DGKZ 6 0.0 FOS 2 0.0 JAK1 16 0.2 CDC25A 14 0.2 RPS6KA1 6 0.0 DCLK1 2 0.0 UBTF 16 0.2 NBN 13 0.2 PALM 6 0.0 TOP2A 2 0.0 HRAS 16 0.2 TP73L 13 0.2 HLA-B 6 0.0 BCL2L1 2 0.0 KHDRBS1 16 0.2 SMN1 13 0.2 RXRA 6 0.0 HNRNPU 2 0.0 CDC2L1 15 0.1 YWHAH 13 0.2 STAT3 6 0.0 ESR2 2 0.0 PSMD10 15 0.1 CDC42 13 0.2 PTPN11 6 0.0 RELA 2 0.0 MED9 15 0.1 TRP73 13 0.2 HIST3H3 6 0.0 SMC1A 2 0.0 NEK6 15 0.1 TGFB1I1 13 0.2 EZH2 6 0.0 HSPA8 2 0.0 SET 15 0.1 SH2D3C 13 0.2 DNMT3B 6 0.0 STMN1 2 0.0 VPS72 15 0.1 FRAP1 13 0.2 TOP2A 6 0.0 EGR1 2 0.0 CSNK2B 15 0.1 SMAD4 13 0.2 POLR2E 6 0.0 CREM 2 0.0 CDC5L 15 0.1 CASP2 13 0.2 CAMK2D 6 0.0 REPS1 2 0.0 TRAF6 14 0.1 RAD9 13 0.2 WWOX 6 0.0 PPP1R12A 2 0.0 ITGB4 14 0.1 MST1R 12 0.1 CDC25C 6 0.0 MYOD1 2 0.0 EZR 14 0.1 PALM 12 0.1 MED10 6 0.0 VDR 1 -0.1 DAPP1 14 0.1 BIRC5 12 0.1 CHKA 6 0.0 POLA1 1 -0.1 MED28 14 0.1 IKBKB 12 0.1 CSNK2A1 6 0.0 ANXA2 1 -0.1 RAD17 14 0.1 HDAC5 12 0.1 GTF2B 6 0.0 AURKB 1 -0.1 SGK1 14 0.1 PIN1 12 0.1 E2 6 0.0 PPP1R1B 1 -0.1 FZR1 14 0.1 IRF3 12 0.1 CDC7 6 0.0 MCM3 1 -0.1 PALM 13 0.1 RIS2 12 0.1 MED28 6 0.0 SPIB 1 -0.1 PCAF 13 0.1 RUNX1 12 0.1 ZBTB17 6 0.0 SLC4A1 1 -0.1 PPM1B 13 0.1 ITGB3 12 0.1 TARS 5 0.0 RNPS1 1 -0.1 PLCG1 13 0.1 RUVBL2 12 0.1 NGFR 5 0.0 ARRB2 1 -0.1 ARID3A 13 0.1 PARP1 12 0.1 ITGB1 5 0.0 BID 1 -0.1 MED10 13 0.1 EPPK1 11 0.1 JAK2 5 0.0 ILK 1 -0.1 PPP2CA 13 0.1 CCNG1 11 0.1 TERF2 5 0.0 KLF1 1 -0.1 PDGFRB 13 0.1 ARID3A 11 0.1 PLK1 5 0.0 EIF2AK2 1 -0.1 TOP2B 13 0.1 SMC1A 11 0.1 KCNK9 5 0.0 TPT1 1 -0.1 APEX1 13 0.1 MRE11A 11 0.1 BAD 5 0.0 GTF2A1 1 -0.1 MDM2 12 0.1 NGFR 11 0.1 GTF2F1 5 0.0 PNKP 1 -0.1 TAF1 12 0.1 IRF7 11 0.1 BCR 5 0.0 JUP 1 -0.1 E2 12 0.1 PRKAR1A 11 0.1 CABLES1 5 0.0 DCLK3 1 -0.1 TERT 12 0.1 CSNK1A1 11 0.1 CDK8 5 0.0 PTPN18 1 -0.1 KRT18 12 0.1 ARRB2 10 0.1 SMARCB1 5 0.0 EXOC4 1 -0.1 ACTB 12 0.1 RGS3 10 0.1 STAT5B 5 0.0 RGS20 1 -0.1 STK11 12 0.1 PAK6 10 0.1 POU2F1 5 0.0 PRKCZ 1 -0.1 CDKN2A 12 0.1 EIF4EBP1 10 0.1 CCNT1 5 0.0 LMNB1 1 -0.1 PIAS1 12 0.1 CHD4 10 0.1 GART 5 0.0 UNC13D 1 -0.1 AXIN1 11 0.1 HSF1 10 0.1 ERCC3 5 0.0 RPS6KA1 1 -0.1 CHEK1 11 0.1 MED9 10 0.1 RUVBL2 5 0.0 EPS8 1 -0.1 DCLK1 11 0.1 THRA 10 0.1 KCNK15 5 0.0 LZTS1 1 -0.1 MAPK10 11 0.1 POLA1 10 0.1 CHD4 5 0.0 COPS5 1 -0.1 THRA 11 0.1 PAICS 10 0.1 EXOS 5 0.0 RBBP4 1 -0.1 SIAH1 11 0.1 ITGB4 9 0.1 ANXA2 4 0.0 CENPB 1 -0.1 VDR 10 0.1 SMARCB1 9 0.1 SUMO1 4 0.0 REV3L 1 -0.1 ROCK1 10 0.1 RHEB 9 0.1 IER3 4 0.0 KRT8 1 -0.1 CCNA1 10 0.1 IRS1 9 0.1 TUBA4A 4 0.0 CTNNA1 1 -0.1 KDR 10 0.1 MED10 9 0.1 PDGFRB 4 0.0 SLK 1 -0.1 SYN1 10 0.1 ADCY6 9 0.1 EPPK1 4 0.0 PFKFB2 1 -0.1 CDK7 10 0.1 PSMC6 9 0.1 MITF 4 0.0 JAK2 1 -0.1 CHEK2 10 0.1 TARS 9 0.1 STK11 4 0.0 ARHGEF7 1 -0.1 CHD4 10 0.1 FGF13 9 0.1 INSR 4 0.0 PDPK1 1 -0.1 TIAM1 10 0.1 CASP9 9 0.1 AKAP13 4 0.0 SNRP70 1 -0.1 BLM 10 0.1 ATXN1 9 0.1 BLM 4 0.0 DMD 1 -0.1 CCND2 10 0.1 NCOA2 9 0.1 PAICS 4 0.0 VPS72 1 -0.1 CABLES1 10 0.1 TJP2 9 0.1 PRAM1 4 0.0 RGS3 1 -0.1 E2F5 10 0.1 FZR1 9 0.1 PCAF 4 0.0 GRK5 1 -0.1 SMARCE1 10 0.1 PNKP 8 0.1 TAF7 4 0.0 AKAP8 1 -0.1 RFC2 9 0.1 TELO2 8 0.1 SLK 4 0.0 BIRC6 1 -0.1 MAP3K5 9 0.1 PA2G4 8 0.1 PFKFB2 4 0.0 POU5F1 1 -0.1 PTK2 9 0.1 LLGL1 8 0.1 PIK3R1 4 0.0 CGN 1 -0.1 ID2 9 0.1 CTBP2 8 0.1 ASCC3L1 4 0.0 HSPA9 1 -0.1 CD5 9 0.1 PHB2 8 0.1 WDR8 4 0.0 ARHGAP1 1 -0.1 CASP9 9 0.1 RAD51 8 0.1 AKAP8 4 0.0 BCR 1 -0.1 CTBP1 9 0.1 HLA-B 8 0.1 MAP3K2 4 0.0 FUSIP1 1 -0.1 CASP4 9 0.1 PIK3R1 8 0.1 ERBB2 4 0.0 TBPL1 1 -0.1 PAX2 9 0.1 VPS72 8 0.1 TOP2B 4 0.0 MVD 1 -0.1 RUVBL2 9 0.1 BAD 8 0.1 CHEK1 4 0.0 ERBB2 1 -0.1 SPNS1 9 0.1 PRKDC 8 0.1 PTP4A3 4 0.0 LIG4 1 -0.1 SRF 9 0.1 LMNA 8 0.1 AATF 4 0.0 GSK3A 1 -0.1 CDK9 8 0.0 CDKN2A 8 0.1 CDC20 4 0.0 NFKBIA 1 -0.1 SMARCB1 8 0.0 PAX2 8 0.1 E2F5 4 0.0 CBX5 1 -0.1 CCNG1 8 0.0 RBP1 8 0.1 PPP2R1A 4 0.0 PTK2 1 -0.1 PAG1 8 0.0 CCNT2 8 0.1 CASP4 4 0.0 E7 1 -0.1 TUBA3C 8 0.0 NCOA6 8 0.1 ATR 4 0.0 PAK6 1 -0.1 HSF1 8 0.0 VDR 7 0.1 BCL2L1 4 0.0 CHEK1 1 -0.1 TUBA1B 8 0.0 IRAK1 7 0.1 DAXX 4 0.0 SFN 1 -0.1 BAD 8 0.0 EIF2AK2 7 0.1 HSP90AB1 4 0.0 PRG2 1 -0.1 FHL5 8 0.0 U2AF2 7 0.1 CHEK2 4 0.0 SNCA 1 -0.1 DBNL 7 0.0 SQSTM1 7 0.1 PIAS1 4 0.0 CSNK1D 1 -0.1 IER2 7 0.0 EIF3A 7 0.1 NCOA6 4 0.0 KHDRBS1 1 -0.1 MCRS1 7 0.0 PRKCI 7 0.1 RET 4 0.0 RHEB 1 -0.1 PPP3R1 7 0.0 DCLK1 7 0.1 AHR 4 0.0 GAPDH 1 -0.1 FOXM1 7 0.0 COIL 7 0.1 E2F2 4 0.0 XRCC1 1 -0.1 FAS 7 0.0 NUMB 7 0.1 FZR1 4 0.0 PRKA4 1 -0.1 SYK 7 0.0 LIN9 7 0.1 HCK 3 0.0 GPI 1 -0.1 CCNT1 7 0.0 EWSR1 7 0.1 ZAP70 3 0.0 ASF1B 1 -0.1 THRB 7 0.0 ITPR1 7 0.1 DUSP12 3 0.0 GNA12 1 -0.1 RPS27A 7 0.0 DNMT3A 7 0.1 PTEN 3 0.0 ERCC3 1 -0.1 ORC3L 7 0.0 UBTF 7 0.1 EWSR1 3 0.0 SPTBN1 1 -0.1 JAG1 7 0.0 HRAS 7 0.1 GRB10 3 0.0 PSEN2 1 -0.1 IKBKE 7 0.0 PHB 7 0.1 PRKCG 3 0.0 STMN2 1 -0.1 LRRC59 7 0.0 KHDRBS1 7 0.1 ZMYND11 3 0.0 SLC9A3R2 1 -0.1 STAT5A 7 0.0 SMARCE1 7 0.1 GTF2I 3 0.0 STK4 1 -0.1 PIM1 7 0.0 CCND3 7 0.1 NCK1 3 0.0 RIS2 1 -0.1 MPP3 7 0.0 RARA 6 0.0 NEDD9 3 0.0 PRDX1 1 -0.1 RAD1 7 0.0 TBK1 6 0.0 JAK1 3 0.0 NUF2 1 -0.1 JAK2 7 0.0 NCOR2 6 0.0 PKD1 3 0.0 AP3B2 1 -0.1 LMNA 7 0.0 RASGRF1 6 0.0 STK6 3 0.0 BRAF 1 -0.1 GART 7 0.0 FADD 6 0.0 PCNA 3 0.0 COPS3 1 -0.1 RANBP9 7 0.0 MAP3K14 6 0.0 NOLC1 3 0.0 VAV3 1 -0.1 VHL 7 0.0 CAPN1 6 0.0 TJP2 3 0.0 GORASP1 1 -0.1 MST1R 6 0.0 MAPK3 6 0.0 XRCC5 3 0.0 MCL1 1 -0.1 SUMO1 6 0.0 DAXX 6 0.0 PRKAB1 3 0.0 CHD4 1 -0.1 RARA 6 0.0 PRKCD 6 0.0 PLCG1 3 0.0 PRKCD 1 -0.1 CD79A 6 0.0 MED28 6 0.0 UBQLN4 3 0.0 SRF 1 -0.1 MEF2A 6 0.0 E4F1 6 0.0 PRKDC 3 0.0 ADCY6 1 -0.1 GPI 6 0.0 ACTB 6 0.0 POLR2F 3 0.0 ID2 1 -0.1 FEN1 6 0.0 YY1 6 0.0 TGFB1I1 3 0.0 PTBP1 1 -0.1 CCNE2 6 0.0 SHANK2 6 0.0 ARID3A 3 0.0 RGS19 1 -0.1 RELA 6 0.0 LYST 6 0.0 SMAD4 3 0.0 CFTR 1 -0.1 RAD51 6 0.0 IKBKG 6 0.0 NPM1 3 0.0 RAD9 1 -0.1 NOLC1 6 0.0 SH3KBP1 6 0.0 ORC6L 3 0.0 PTPN2 1 -0.1 CUX1 6 0.0 RPS6KA1 6 0.0 SLC9A3R2 3 0.0 RPS6KA3 1 -0.1 GRIN2A 6 0.0 SGK1 6 0.0 ORC3L 3 0.0 PKN1 1 -0.1 NF2 6 0.0 JAK2 6 0.0 VCP 3 0.0 TFAP2A 1 -0.1 APP 6 0.0 CHUK 6 0.0 PSMD10 3 0.0 HNRNPC 1 -0.1 CSF2RB 6 0.0 PFTK1 6 0.0 CDC6 3 0.0 CRMP1 1 -0.1 PARP1 6 0.0 ARRB1 6 0.0 SMC1A 3 0.0 CDC37 1 -0.1 DUSP19 5 0.0 STAT1 6 0.0 VAV1 3 0.0 ZFYVE9 1 -0.1 FCGR2A 5 0.0 MAST1 6 0.0 STAT5A 3 0.0 ERCC6 1 -0.1 MDC1 5 0.0 DCLRE1C 6 0.0 MTA1 3 0.0 PDE3B 1 -0.1 RB1CC1 5 0.0 YWHAB 6 0.0 PSMC6 3 0.0 RAD17 1 -0.1 SMARCC1 5 0.0 SUMO1 5 0.0 ITGB3BP 3 0.0 SIAH1 1 -0.1 FLCN 5 0.0 HIST3H3 5 0.0 RAD17 3 0.0 YWHAZ 1 -0.1 STX1A 5 0.0 SMARCA4 5 0.0 ACTB 2 0.0 SYNJ2BP 1 -0.1 WDR8 5 0.0 LRP2 5 0.0 CD44 2 0.0 CDC25A 1 -0.1 AKAP1 5 0.0 CDC2L6 5 0.0 PLSCR4 2 0.0 EDD 1 -0.1 MED17 5 0.0 TERT 5 0.0 MTHFD1 2 0.0 PKN2 1 -0.1 PLEC1 5 0.0 IKBKE 5 0.0 GTF2F2 2 0.0 CASP10 5 0.0 MED14 5 0.0 CTNND2 2 0.0 STAT5B 5 0.0 MYBL2 5 0.0 CDK6 2 0.0 CDC2L6 5 0.0 SLC9A3R1 5 0.0 MAST2 2 0.0 FBL 5 0.0 WRN 5 0.0 AXIN1 2 0.0 LIN9 5 0.0 NEK6 5 0.0 ID3 2 0.0 PA2G4 5 0.0 SP3 5 0.0 CXORF53 2 0.0 NUAK1 5 0.0 PDGFRB 5 0.0 CDC2L5 2 0.0 DOK4 5 0.0 TH 5 0.0 SLC6A4 2 0.0 IER3 5 0.0 ZMYND11 5 0.0 SHC 2 0.0 CDK6 5 0.0 INSR 5 0.0 TUBA1A 2 0.0 PHB2 5 0.0 GANAB 5 0.0 PLCG2 2 0.0 IKBKG 5 0.0 RBBP4 5 0.0 VIM 2 0.0 ATXN1 5 0.0 MED26 5 0.0 BIRC4 2 0.0 PDPK1 5 0.0 PTP4A3 5 0.0 SMURF1 2 0.0 SMN2 5 0.0 GEMIN4 5 0.0 STX1A 2 0.0 BMPR1B 5 0.0 VDAC1 5 0.0 PIAS4 2 0.0 SMARCA2 5 0.0 NKX2-1 5 0.0 NME2 2 0.0 CALD1 5 0.0 MAP3K1 5 0.0 CHUK 2 0.0 CD79B 5 0.0 KIF5C 5 0.0 BARD1 2 0.0 APC 5 0.0 ITGB5 5 0.0 MDM2 2 0.0 ETV6 5 0.0 STMN1 5 0.0 HSPA9 2 0.0 MED22 5 0.0 YWHAG 4 0.0 CCND1 2 0.0 CALM1 5 0.0 MARCKS 4 0.0 POLE2 2 0.0 FLNA 5 0.0 HMGN1 4 0.0 BTRC 2 0.0 STMN1 5 0.0 SNIP1 4 0.0 UBTF 2 0.0 RYBP 5 0.0 MAGI2 4 0.0 VDAC1 2 0.0 HNRNPC 5 0.0 UBQLN4 4 0.0 CSNK1D 2 0.0 ANXA2 4 0.0 SKP2 4 0.0 PHB 2 0.0 YBX1 4 0.0 GPI 4 0.0 TAF4 2 0.0 DUSP12 4 0.0 GNB2L1 4 0.0 DYNLL1 2 0.0 WASL 4 0.0 NRIP1 4 0.0 ELF3 2 0.0 CLSTN1 4 0.0 MAPK10 4 0.0 MAPK8 2 0.0 FGFR1 4 0.0 CDX2 4 0.0 TAF10 2 0.0 PLD2 4 0.0 NCOA1 4 0.0 RBP1 2 0.0 PKD1 4 0.0 KRT18 4 0.0 TREX1 2 0.0 INTS6 4 0.0 PPP2CA 4 0.0 TSC22D3 2 0.0 GTF2F1 4 0.0 AURKB 4 0.0 CALM1 2 0.0 BCR 4 0.0 NUAK1 4 0.0 PRDX1 2 0.0 PRG2 4 0.0 PLSCR4 4 0.0 FLNA 2 0.0 SH2B2 4 0.0 RNPS1 4 0.0 EIF4EBP1 2 0.0 BCL2L1 4 0.0 RAD1 4 0.0 PPP1R9B 2 0.0 ETS1 4 0.0 HTRA2 4 0.0 CDK7 2 0.0 NCOA1 4 0.0 PELI3 4 0.0 TNFAIP3 2 0.0 SMC1A 4 0.0 KRT8 4 0.0 CIITA 2 0.0 MCC 4 0.0 SERTAD1 4 0.0 TERT 2 0.0 CDKN1C 4 0.0 XRCC5 4 0.0 IKBKE 2 0.0 MED31 4 0.0 ELF4 4 0.0 TRRAP 2 0.0 SP3 4 0.0 CABLES1 4 0.0 KRTHA3A 2 0.0 NCOA3 4 0.0 ATF5 4 0.0 RANBP9 2 0.0 INSR 4 0.0 CASP7 4 0.0 COBRA1 2 0.0 SIRT2 4 0.0 E2F5 4 0.0 SMARCE1 2 0.0 MARK4 4 0.0 ETV6 4 0.0 MYBL2 2 0.0 CIB1 4 0.0 APP 4 0.0 PPIG 2 0.0 NEDD9 4 0.0 TRADD 4 0.0 ID2 2 0.0 TERF2 4 0.0 PRKCSH 3 0.0 BATF 2 0.0 AKAP8 4 0.0 BID 3 0.0 MAPK9 2 0.0 PCTK1 4 0.0 RAC1 3 0.0 HTATSF1 2 0.0 PFTK1 4 0.0 OGT 3 0.0 ORC5L 2 0.0 ACTA1 4 0.0 AXIN1 3 0.0 NFYA 2 0.0 GSK3A 4 0.0 CCNB1 3 0.0 WEE1 2 0.0 DCC 4 0.0 TOPBP1 3 0.0 POLR2H 2 0.0 ORC6L 4 0.0 POLI 3 0.0 RYBP 2 0.0 ERCC3 4 0.0 PTPN1 3 0.0 E4F1 2 0.0 CRKL 4 0.0 BRAP 3 0.0 BRE 2 0.0 NKX2-1 4 0.0 KCNK3 3 0.0 POLE 2 0.0 VCP 4 0.0 ING3 3 0.0 CLK1 2 0.0 E2F3 4 0.0 NS5A 3 0.0 HSPH1 2 0.0 GAB2 4 0.0 GTF2H1 3 0.0 BRF1 2 0.0 TRIP10 4 0.0 NME2 3 0.0 PRPF40A 1 -0.1 CTDP1 4 0.0 SH3BP2 3 0.0 YBX1 1 -0.1 HDAC6 4 0.0 NFATC1 3 0.0 GPHN 1 -0.1 BIN1 4 0.0 WDR8 3 0.0 NUAK1 1 -0.1 PINK1 3 0.0 SFRS1 3 0.0 SKP1A 1 -0.1 OGT 3 0.0 KCNK9 3 0.0 NBN 1 -0.1 NTRK1 3 0.0 HSPA9 3 0.0 IER2 1 -0.1 MITF 3 0.0 NR4A1 3 0.0 SMN1 1 -0.1 PTPN18 3 0.0 BTRC 3 0.0 KDR 1 -0.1 SMARCC2 3 0.0 FOS 3 0.0 CLSPN 1 -0.1 DAB2 3 0.0 IQGAP1 3 0.0 COPB1 1 -0.1 PTPRE 3 0.0 CAMK2A 3 0.0 CLSTN1 1 -0.1 PTK2B 3 0.0 PDE6D 3 0.0 JUND 1 -0.1 TELO2 3 0.0 EEF1E1 3 0.0 PHB2 1 -0.1 PPP5C 3 0.0 SYN1 3 0.0 GANAB 1 -0.1 CLN5 3 0.0 SPTBN1 3 0.0 RARA 1 -0.1 NEDD4 3 0.0 EZR 3 0.0 POLI 1 -0.1 FHOD1 3 0.0 DYNC1H1 3 0.0 RAD51 1 -0.1 DNM1 3 0.0 KCNK15 3 0.0 CD4 1 -0.1 SETDB1 3 0.0 CDC6 3 0.0 ING3 1 -0.1 CAMK2A 3 0.0 PPARG 3 0.0 SH3KBP1 1 -0.1 PTHLH 3 0.0 JAG1 3 0.0 TLK1 1 -0.1 PPP2R1A 3 0.0 FBL 3 0.0 CUL7 1 -0.1 ACVR1 3 0.0 TIAM1 3 0.0 HMGN1 1 -0.1 COIL 3 0.0 RUNX2 3 0.0 ADRA1B 1 -0.1 TSC2 3 0.0 STUB1 3 0.0 SNCAIP 1 -0.1 RUNX1 3 0.0 NCOA3 3 0.0 NONO 1 -0.1 CFTR 3 0.0 MTHFD1 3 0.0 GTF2H1 1 -0.1 TFAP2A 3 0.0 CDK6 3 0.0 SET 1 -0.1 PPP1R12A 3 0.0 CSK 3 0.0 PRKAR1A 1 -0.1 ATP1A1 3 0.0 SMAD7 3 0.0 RUNX3 1 -0.1 YWHAZ 3 0.0 ZBTB16 3 0.0 NASP 1 -0.1 KLKB1 3 0.0 MAPK8IP3 3 0.0 MAF 1 -0.1 PPP1R1B 3 0.0 CSNK2B 3 0.0 NFATC1 1 -0.1 CDKN2D 3 0.0 CUX1 3 0.0 POU5F1 1 -0.1 ILK 3 0.0 PCTK1 3 0.0 GNAS 1 -0.1 TGIF1 3 0.0 TGFBR1 3 0.0 NEDD4 1 -0.1 CLSPN 3 0.0 LIG4 3 0.0 MAP2K1 1 -0.1 VASP 3 0.0 GSK3A 3 0.0 NFKB1 1 -0.1 ZYX 3 0.0 REM1 3 0.0 ATXN7 1 -0.1 PELI3 3 0.0 CASP4 3 0.0 IKBKB 1 -0.1 WAS 3 0.0 APEX1 3 0.0 GSPT2 1 -0.1 ARRB1 3 0.0 MAPK8IP2 3 0.0 ELF4 1 -0.1 GIPC1 3 0.0 GNA13 3 0.0 MED26 1 -0.1 ITK 3 0.0 VAV1 3 0.0 GTF3C3 1 -0.1 IFNAR2 3 0.0 PRKCB1 3 0.0 ORC4L 1 -0.1 ING1 3 0.0 EXOS 3 0.0 RBX1 1 -0.1 SUPT16H 3 0.0 TOP1 3 0.0 NR2E3 1 -0.1 TREX1 3 0.0 MYB 3 0.0 PELO 1 -0.1 BRAF 3 0.0 ANXA2 2 0.0 CREB1 1 -0.1 GPS2 3 0.0 PRKCE 2 0.0 PPP2R3B 1 -0.1 BATF 3 0.0 DLG1 2 0.0 GRB2 1 -0.1 EEF1A1 3 0.0 PINK1 2 0.0 CDK4 1 -0.1 SMARCD3 2 0.0 ROCK1 2 0.0 TERF1 1 -0.1 ZAP70 2 0.0 DDEF2 2 0.0 FKBP3 1 -0.1 POLR2G 2 0.0 KLF1 2 0.0 ING1 1 -0.1 NBN 2 0.0 HNRPD 2 0.0 ASF1B 1 -0.1 EIF2AK2 2 0.0 NTRK1 2 0.0 RBBP8 1 -0.1 RAC1 2 0.0 MITF 2 0.0 SPTBN1 1 -0.1 PRKAR2B 2 0.0 EXOC4 2 0.0 SND1 1 -0.1 PTEN 2 0.0 LMNB1 2 0.0 IRS1 1 -0.1 U2AF2 2 0.0 KIAA1446 2 0.0 RNASEH2B 1 -0.1 MLLT7 2 0.0 CDK9 2 0.0 ETV1 1 -0.1 UBA1 2 0.0 RXRA 2 0.0 KIAA1967 1 -0.1 GTF2H1 2 0.0 MAP3K5 2 0.0 ETS2 1 -0.1 BIRC4 2 0.0 NASP 2 0.0 NFE2L1 1 -0.1 EBNA1 2 0.0 NOS1 2 0.0 AP3B2 1 -0.1 SLK 2 0.0 PRMT5 2 0.0 GSPT1 1 -0.1 PFKFB2 2 0.0 KPNA2 2 0.0 BAX 1 -0.1 SNIP1 2 0.0 PRKG1 2 0.0 GPS2 1 -0.1 SQSTM1 2 0.0 GNAS 2 0.0 JAG1 1 -0.1 BIRC5 2 0.0 NPHP1 2 0.0 TRIP10 1 -0.1 IKBKB 2 0.0 GTF2F1 2 0.0 HIST1H1A 1 -0.1 ARHGAP1 2 0.0 PCYT1A 2 0.0 TGFA 1 -0.1 PCYT1A 2 0.0 BCR 2 0.0 SFRS3 1 -0.1 PURA 2 0.0 CASP10 2 0.0 DBF4 1 -0.1 HAND1 2 0.0 PELO 2 0.0 EGR1 1 -0.1 FOS 2 0.0 FAS 2 0.0 CREM 1 -0.1 TOM1L1 2 0.0 HAND1 2 0.0 EEF1A1 1 -0.1 CRIP2 2 0.0 TUBB 2 0.0 TIAM1 1 -0.1 H3F3B 2 0.0 HOOK2 2 0.0 NCKIPSD 1 -0.1 TAF11 2 0.0 STAT5B 2 0.0 PKN1 1 -0.1 HSP90AB1 2 0.0 PIK3CB 2 0.0 TFAP2A 1 -0.1 EIF4EBP1 2 0.0 SUPT5H 2 0.0 BIN1 1 -0.1 PPARG 2 0.0 TCERG1 2 0.0 HMGB1 1 -0.1 CIITA 2 0.0 MEF2A 2 0.0 PARP1 1 -0.1 TRRAP 2 0.0 FEN1 2 0.0 NR3C1 1 -0.1 ADCY6 2 0.0 SND1 2 0.0 TOP1 1 -0.1 HCFC1 2 0.0 APLP2 2 0.0 NCKAP1 1 -0.1 SORBS2 2 0.0 SDC2 2 0.0 THRA 1 -0.1 INPP5D 2 0.0 ETS2 2 0.0 HUWE1 1 -0.1 HSPH1 2 0.0 NFE2L1 2 0.0 HDAC3 2 0.0 ARPC1B 2 0.0 IL7R 2 0.0 MARK3 2 0.0 MAP4K1 2 0.0 HIST1H1A 2 0.0 IL16 2 0.0 TNFRSF14 2 0.0 PITPNA 2 0.0 RGS19 2 0.0 RNPS1 2 0.0 TRIM37 2 0.0 MAP1B 2 0.0 FBXW7 2 0.0 HIST4H4 2 0.0 KLKB1 2 0.0 RANBP10 2 0.0 POLE 2 0.0 MORF4L2 2 0.0 GFAP 2 0.0 ACP1 2 0.0 MAP4K1 2 0.0 SYNJ1 2 0.0 ILK 2 0.0 BRD2 2 0.0 MAP1B 2 0.0 ZBTB16 2 0.0 TPT1 2 0.0 VIM 2 0.0 CENPJ 2 0.0 NONO 2 0.0 IER3 2 0.0 RBBP4 2 0.0 NGFRAP1 2 0.0 CDKN2C 2 0.0 CIB1 2 0.0 CHUK 2 0.0 NEDD9 2 0.0 RUVBL1 2 0.0 NONO 2 0.0 XRCC5 2 0.0 NOLC1 2 0.0 PAK3 2 0.0 CASP1 2 0.0 ELF4 2 0.0 PPFIBP1 2 0.0 MED26 2 0.0 EIF3K 2 0.0 PCGF2 2 0.0 NCF1 2 0.0 GEMIN4 2 0.0 NFKB1 2 0.0 AATF 2 0.0 HSPA1A 2 0.0 FOXO3 2 0.0 IRS2 2 0.0 GAPDH 2 0.0 PSEN1 2 0.0 PSEN1 2 0.0 ING1 2 0.0 STMN2 2 0.0 TP53BP1 2 0.0 PARK2 2 0.0 ERCC3 2 0.0 PLAUR 2 0.0 ATP6V1E1 2 0.0 CEBPE 2 0.0 STMN2 2 0.0 MAP3K1 2 0.0 ETV1 2 0.0 ARID2 2 0.0 NF2 2 0.0 AP3B2 2 0.0 TREX1 2 0.0 CTTN 2 0.0 WWOX 2 0.0 TRADD 2 0.0 CALM1 2 0.0 RPS6KA2 2 0.0 CEBPE 2 0.0 PDE3B 2 0.0 VCP 2 0.0 SYP 2 0.0 AP3B2 2 0.0 TOP1 2 0.0 GORASP1 2 0.0 PTPRC 2 0.0 MCL1 2 0.0 MYB 2 0.0 CAMK2G 2 0.0 SKP1A 1 -0.1 TP53BP2 2 0.0 KLF1 1 -0.1 BATF 2 0.0 HNRPD 1 -0.1 FGFBP1 2 0.0 PNKP 1 -0.1 PTPRC 2 0.0 DAF 1 -0.1 YBX1 1 -0.1 EXOC4 1 -0.1 ZAP70 1 -0.1 USF2 1 -0.1 POLR2G 1 -0.1 CSNK1E 1 -0.1 PSTPIP1 1 -0.1 YEATS4 1 -0.1 SKP1A 1 -0.1 NR2F1 1 -0.1 SOS1 1 -0.1 RPS6KA5 1 -0.1 SRCAP 1 -0.1 MAPKAPK2 1 -0.1 TUBA4A 1 -0.1 STK6 1 -0.1 SMARCC2 1 -0.1 RFC3 1 -0.1 SMARCC1 1 -0.1 AFAP1L2 1 -0.1 USF2 1 -0.1 DIAPH2 1 -0.1 MLLT7 1 -0.1 TBK1 1 -0.1 BAK1 1 -0.1 BAZ1B 1 -0.1 BBC3 1 -0.1 TXK 1 -0.1 PLD2 1 -0.1 NFATC1 1 -0.1 ARF6 1 -0.1 SIC1 1 -0.1 PLCG2 1 -0.1 DNMT1 1 -0.1 BIRC4 1 -0.1 POLR1B 1 -0.1 STK6 1 -0.1 PLK3 1 -0.1 RFC3 1 -0.1 SUV39H1 1 -0.1 STX1A 1 -0.1 PAX6 1 -0.1 GDNF 1 -0.1 MKNK1 1 -0.1 AFAP1L2 1 -0.1 PAFAH1B1 1 -0.1 TGM2 1 -0.1 CCNB2 1 -0.1 RUNX3 1 -0.1 RASGRF1 1 -0.1 SP100 1 -0.1 FADD 1 -0.1 HRK 1 -0.1 ERBB2 1 -0.1 DNMT1 1 -0.1 BTRC 1 -0.1 AKAP1 1 -0.1 IQGAP1 1 -0.1 PPP3R1 1 -0.1 YWHAQ 1 -0.1 PPM1B 1 -0.1 XPA 1 -0.1 SUV39H1 1 -0.1 KIAA1377 1 -0.1 PPP3CA 1 -0.1 MAP3K14 1 -0.1 MED17 1 -0.1 PPP1R9A 1 -0.1 TUBA8 1 -0.1 SKI 1 -0.1 ARHGAP1 1 -0.1 AP1B1 1 -0.1 PAFAH1B1 1 -0.1 SUPT5H 1 -0.1 PLEC1 1 -0.1 TRPV4 1 -0.1 TAF1 1 -0.1 KCNH2 1 -0.1 FOXM1 1 -0.1 CAMK2D 1 -0.1 PF4 1 -0.1 SPTBN1 1 -0.1 DAPK3 1 -0.1 SND1 1 -0.1 PTK2 1 -0.1 MED8 1 -0.1 XPO1 1 -0.1 ACTC1 1 -0.1 SETDB1 1 -0.1 DYNC1H1 1 -0.1 XPA 1 -0.1 PPP1R9B 1 -0.1 KIAA1377 1 -0.1 BAG1 1 -0.1 PTHLH 1 -0.1 TFAP2B 1 -0.1 CCNT1 1 -0.1 DNM2 1 -0.1 ERBB2IP 1 -0.1 MED21 1 -0.1 KLF5 1 -0.1 BRD8 1 -0.1 KCNH2 1 -0.1 MRE11A 1 -0.1 RBBP8 1 -0.1 STAU1 1 -0.1 LUC7L2 1 -0.1 RGS19 1 -0.1 CAMK2D 1 -0.1 HTATSF1 1 -0.1 RNASEH2B 1 -0.1 ZEB2 1 -0.1 TUBA3C 1 -0.1 NCKIPSD 1 -0.1 TSG101 1 -0.1 FBXW7 1 -0.1 RBBP7 1 -0.1 MED15 1 -0.1 TSC22D3 1 -0.1 HEXIM1 1 -0.1 APOB 1 -0.1 E2F2 1 -0.1 HSP90AB1 1 -0.1 BRF1 1 -0.1 ZRSR2 1 -0.1 HUWE1 1 -0.1 BCAS1 1 -0.1 BTK 1 -0.1 CDK7 1 -0.1 AURKB 1 -0.1 BAG1 1 -0.1 FAF1 1 -0.1 TFAP2B 1 -0.1 YWHAH 1 -0.1 CFTR 1 -0.1 EWSR1 1 -0.1 STAT5A 1 -0.1 COPB1 1 -0.1 DBF4 1 -0.1 JUND 1 -0.1 EGR1 1 -0.1 SUB1 1 -0.1 MAP3K4 1 -0.1 SMARCA5 1 -0.1 CYLD 1 -0.1 SH3KBP1 1 -0.1 POLR2H 1 -0.1 HTRA2 1 -0.1 TUBA1B 1 -0.1 SMAD7 1 -0.1 CDR2 1 -0.1 TANK 1 -0.1 MTA1 1 -0.1 MYCBP 1 -0.1 ATP1A1 1 -0.1 FHIT 1 -0.1 PIM1 1 -0.1 MPP1 1 -0.1 HUWE1 1 -0.1 APAF1 1 -0.1 BRF1 1 -0.1 GRAP 1 -0.1 HDAC3 1 -0.1 NMT1 1 -0.1 BTK 1 -0.1 YAP1 1 -0.1 PPP1CC 1 -0.1 ARR3 1 -0.1 PAXIP1 1 -0.1 BHLHB2 1 -0.1 UIMC1 1 -0.1 CGN 1 -0.1 DISC1 1 -0.1 TPM1 1 -0.1 LEF1 1 -0.1 MAPKBP1 1 -0.1 DOK4 1 -0.1 PSMC3 1 -0.1 SRFBP1 1 -0.1 ORC4L 1 -0.1 PACSIN3 1 -0.1 E2F4 1 -0.1 MORF4L2 1 -0.1 NSMAF 1 -0.1 ASAP1 1 -0.1 NUMA1 1 -0.1 SYNJ1 1 -0.1 SORBS3 1 -0.1 TCF7L2 1 -0.1 DMTF1 1 -0.1 JAK1 1 -0.1 SFN 1 -0.1 ADAM17 1 -0.1 ACTN1 1 -0.1 SET 1 -0.1 PHB 1 -0.1 HIPK2 1 -0.1 ITGB2 1 -0.1 RAD52 1 -0.1 POU2F1 1 -0.1 CCND2 1 -0.1 PRKA4 1 -0.1 FHIT 1 -0.1 HIST3H2BB 1 -0.1 MLXIPL 1 -0.1 EEF2K 1 -0.1 GCN5L2 1 -0.1 CASP2 1 -0.1 MAPKBP1 1 -0.1 PSEN2 1 -0.1 TDRD7 1 -0.1 RANBP2 1 -0.1 SMARCA2 1 -0.1 BIRC2 1 -0.1 BMX 1 -0.1 RBP1 1 -0.1 CALD1 1 -0.1 SLC9A3R2 1 -0.1 RRAS 1 -0.1 WWOX 1 -0.1 RPL31 1 -0.1 EPHB2 1 -0.1 PCGF2 1 -0.1 CD3E 1 -0.1 CDC20 1 -0.1 ACHE 1 -0.1 POU2F1 1 -0.1 PTPN13 1 -0.1 GNA12 1 -0.1 GMEB1 1 -0.1 MED22 1 -0.1 FOXO1 1 -0.1 EEF2K 1 -0.1 TESK1 1 -0.1 HOXB7 1 -0.1 FBXO7 1 -0.1 CEP57 1 -0.1 CCND3 1 -0.1 BRAF 1 -0.1 SMYD2 1 -0.1 PAX5 1 -0.1 MED30 1 -0.1 LASP1 1 -0.1 ORC5L 1 -0.1 BAX 1 -0.1 LIMK2 1 -0.1 PTPN13 1 -0.1 DCLRE1C 1 -0.1 HGS 1 -0.1 TNNT1 1 -0.1 ACTN2 1 -0.1 RBAK 1 -0.1 LNX1 1 -0.1 CASP8AP2 1 -0.1 RYBP 1 -0.1 ITGB3BP 1 -0.1 BIN1 1 -0.1

Table S6: Occurrences of nodes in shortest path networks. Table S6 gives the complete list of molecules and the number of times each molecules occurred in shortest paths networks along with the normalized z-score for each molecule. Table S8

Classification of nodes according to their presence in different phase specific modules.

Common Common Molecules Functional molecules Molecules Unique to molecules Molecules unique to groups between G1 G1S between unique to S G1 and G1S G1S and S

YWHAQ ADRB2 RGS16 ABL1 CDK2 BAD AKT1 CHEK2 CDK5 HRAS CDC25B WASL EGFR CRK PDGFRB GNB2L1 JAK1 LCK GRB2 PRKCD LYN PTK2B CSNK2A1 MNAT1 PTPN2 MAPK8 ERBB2 MAPK14 PTPN6 PTPN12 MST1R CDKN1B PTPRE Signaling SYNGAP1 PDE6G YWHAB RPS6KA1 intermediates ASAP1 PDPK1 PELP1 CASP9 PRG2 PTPN18 PRKDC PIAS3 ARHGEF7 SHB MAPK3 JAK2 SHC MAP2K1 NBN SRC MDM2 CTNNB1 BCAR1

CREBBP SP1 TRIP4 BRCA1 TAF1 E2F1 ELK1 TRP73 E7 ESR1 EP300 STAT5A FOXO1 GTF2I CITED1 Transcription FOXO3 JUN SMAD2 Regulators IKBKB MYB SMARCA4 MYC STAT1 SMAD3 STAT3 STAT5B E6 TRP53 NCOA3 TSC2

SKP2 YTHDC1 RIS2 DNA replication SP3 MCM7 factors

HSP90AA1 CCNT1 RB1 CCNA2 RBL2 Other YTHDC1 RBP1 RBL1 HDAC1 Cell Cycle Regulators LEGENDS FOR SUPPLEMENTARY TABLES

Table S1: siRNA screen results for targeted kinases and phosphatases. Table S1 gives the complete details of the primary and validation screens of all kinases and phosphatases. This includes siRNA target sequences, the raw value for cell cycle distribution obtained from FACS, and the frequency of dead cell populations. Data for each of the replicates, with their normalized values, are provided along with the final mean of two replicates used for hit selection. The table gives additional information about ontology of the molecules and variants. Table S2: Gene expression status of the validated hits. Table S2 provides the expression status of the final validated hits in CH1 cells. Raw expression values for the 43 hits identified by our RNAi screen are listed with corresponding probe sequence. Microarray data are available in the ArrayExpression Database (Accession No. E-MTAB-82).

Table S5: High confidence network used for graph theory analysis. The complete interaction file compiled by merging and filtering the different databases have been provided in table S5 as a two column file.

Table S7: Network file used as SNAVI background. The flat file used for detection of motifs using SNAVI is provided in table S7 with the complete information regarding the interactors, the type of interaction, biochemical process and the PMID providing evidences for the interaction.

SUPPLEMENTARY EXPERIMENTAL PROCEDURES.

Cell Lines and Cell Culture Medium

Murine B cell Lymphoma CH1 (TIB-221) cells were cultured in RPMI 1640 (Invitrogen)

medium with 2mM L-glutamine adjusted to contain 1.5 g/L Na bicarbonate, 4.5 g/L glucose,

10 mM HEPES, 0.05 mM 2-mercaptoethanol (90%) ,1.0 mM Na pyruvate and 10% fetal calf

serum.

Human cell lines Raji, Jurkat, HL60, U937 and THP1 were cultured in RPMI 1640 medium

with 2mM L-glutamine adjusted to contain 1.5g/L Na bicarbonate,4.5g/L glucose,10 mM

HEPES, and 10% fetal calf serum.

Human Burkitts Lymphoma Namalwa was cultured in RPMI 1640 medium with 2mM L-

glutamine adjusted to contain 1.5g/L Na bicarbonate,4.5g/L glucose,10 mM HEPES, and 1.0

mM Na pyruvate and 10% fetal calf serum.

Human cell lines A549, HeLa, HepG2, HEK293, U2OS, AW8507, MCF7 and Huh7 and

were maintained in DMEM (Dulbecco’s Modified Eagle Medium, Invitrogen) supplemented

with 10% FCS. All cell lines were cultures at 370C and 5% CO2.

siRNA libraries

The primary siRNA screens were performed using siRNA libraries targeting mouse Kinases

(Mouse Kinase siRNA SET V1.0) and Phosphatases (Mouse Phosphatase siRNA SET V1.0)

containing 2 siRNA duplexes for each target gene (Qiagen).The siRNAs were designed using

the innovative HiPerformance siRNA design algorithm. A pool of the 2 siRNAs was used and a total of 758 Kinases and 294 phosphatases were screened.

The Validation screen for the shortlisted Kinase hits was performed using pools of 3

individual siRNA duplexes against each gene target from the MISSION siRNA Mouse Kinase Panel (Sigma).The screen was performed similar to the primary screen. For

Validation of Phosphatases, siGENOME (Dhramacon) smartpool siRNAs containing 4 siRNAs per target gene were used.

Standardization of the screening procedure:

(i) No. of Cells Plated/ Well

Experiments were done to determine the optimum number of cells to be used per well for transfection and subsequently for cell cycle analysis. CH1 cells were seeded at varying numbers ranging from 10,000/well to 50,000/well and transfected with GFP-specific siRNA.

Subsequently, cell viability was determined using trypan blue exclusion at 24, 48, and 72 hours. A seed density of 20,000 cells/well was found to be optimal with a cell viability that was consistently >85%.

(ii) siRNA Transfection Conditions

Transfection efficiency: To ensure efficient transfection Alexa 488 labelled siRNA was used to transfect CH1 cells. 20,000 cells were plated in a 96-well plate and transfected using

HiPerfect transfection reagent. Transfection was carried out as per the manufacturer’s protocol. The efficiency was monitored at up to 72 hours post transfection using confocal microscopy and was found to be >98%.

Optimal siRNA concentration and the kinetics of Knockdown: To determine optimal knockdown conditions siRNAs against six representative proteins were chosen. Using the optimised cell number (20,000/well) experiments were set up with different concentrations of siRNA ranging from 50 to 400nM using HiPerfect transfection reagent as per manufacturers protocol. The effect of different siRNA knockdowns on protein level was determined by

Western blot analysis in cell lysates obtained at 24, 36, 48, 72 and 96h after transfection.

Across experiments the siRNA concentration of 200nM was found to be optimal, giving a knockdown of between 70-80%, which was retained up to 96h post transfection. The western

blots for two representative molecules BTK and PLK1, along with those for three validated

screen targets are shown in Figure S1B. The relatively higher concentration of siRNA

required is likely due to the fact that these cells grow as a suspension.

siRNA Screen

The mouse Kinase and phosphatase libraries (Qiagen) containing 2 siRNAs per target gene

were obtained in a 96 well format. The siRNAs were diluted to a working concentration of

10µM. CH1 mouse B cell lymphoma cells were transfected with the pool of 2 siRNA

duplexes per target gene, using HiPerfect transfection reagent as per the manufacturer’s

protocol for transfection of suspension cell lines (Qiagen). Briefly, 20,000 cells of mouse

lymphoma cell line CH1 in 25uL of RPMI were seeded into each well of a flat bottom 96

0 well plate. The cells were incubated at 37 C and 5% CO2 for 30 minutes. Meanwhile, 25uL

(per well) of transfection mix, containing RPMI medium and a 1:1 volume ratio of siRNA

and transfection reagent Hiperfect i.e. 200nM (3ul) siRNA and 3ul HiPerfect were prepared

in parallel U-bottom 96 well plates. To each well in the plate containing cells, 25ul of

transfection mix was added. The mixture was pipetted a few times to ensure proper mixing of

0 transfection complexes with the cells. The plates were incubated at 37 C and 5% CO2. At 6

hours post transfection the volume of each well was made up to 150 uL with complete

medium (RPM1 with 10% FCS, 1mM sodium Pyruvate, Bme). At 24 and 48 hours post

transfection the medium of the plate was replaced with fresh complete medium. Our

transfection conditions ensured that >98% of the cells were transfected with the siRNA.

Further, by using siRNA pools against a representative set of cellular kinases we determined

that specific depletion was evident by 36 h after transfection, and that this attained maximal

levels by 48h. Depleted protein levels then persisted up to 96h of the culture (Figure S1B). Since during standardization experiments the target protein levels were found to be the lowest

at 48 hours post transfection, the cell cycle analysis was performed at 72 hours (which is 24

hours after the point at which protein levels are minimum). Each plate had 5 negative control

wells which included cells treated with scrambled siRNA (2 wells), GFP siRNA (2 wells) and

1 well contained mock transfected cells (i.e. only HiPerfect). All the screens were performed

in duplicates.

Propidium Iodide staining and sample acquisition

For propidium iodide staining and DNA content analysis the cells were centrifuged in the 96

well plates using a plate rotor and the medium was aspirated. The cells were stained for 30

minutes at 40C with 150 uL propidium iodide staining solution containing 0.1mg/mL

Propidium Iodide (Sigma), 3ul/mL TritonX -100 (Sigma),1 mg/mL sodium citrate (Sigma) and 20ug/mL RNase (Sigma). Plates which were not acquired immediately were kept at 40C.

The samples were acquired in a 96-well plate format using an automated BD FACSCalibur

HTS or High Throughput Sampler (Beckton Dickinson, FACSCalibur).

Primary and secondary screen data analysis and hit selection

The data obtained was analysed using FlowJo software and the DNA histograms obtained

were analysed to quantitate the subG1, G1, S and G2 populations. During data analysis we observed minor shifts in G1 and G2 peak positions and widths across cell cycle histograms.

This prevented a reliable quantification and comparison of S and G2 phases of the cell cycle

using automated platforms such as those available with FlowJo. The cell cycle histograms

were therefore analysed using manual gating of the sub-G1, G1, S and G2 phases. The G1, S

and G2 phases were analysed after gating out the subG1 population such that their quantified

values represent percentages of the total live cell population. For each plate the gates were defined for the negative control samples and the same gates were copied to the other histograms and used for analysis. Although with respect to the control sample there were some well-dependent (instrument based) and sample dependent shifts in peaks of some samples in each plate. To increase precision of analysis small adjustments in the gates of individual peaks of each sample allowed reliable quantification of cell cycle phases. The same procedure was used for the analysis of the data obtained from the second screen.

To allow the comparison of data sets from different plates the data was normalized by calculating and representing each quantified parameter sub G1, G1, S and G2 populations into Z scores (also called normal scores). Each Z-score is a dimensionless quantity which depicts the deviation of the data point from the mean of the negative controls in units of the standard deviations of the negative control. The Z scores were calculated for each parameter with respect to the mean and standard deviations of the 5 negative controls of its respective plate. Each Z score was calculated as follows:

Z score=(x-µ)/σ

Where x is the G1/S/G2 population to be standardised, µ is the plate mean/mean of controls and σ is the standard deviation of the plate/plate controls. Both the screens were performed in duplicates, therefore for each siRNA knockdown there were two Z scores. The Z scores were obtained for each duplicate and the mean was calculated. For the first screen a z score cut off of Z > ± 2.5 was chosen and 83 kinases and 22 phosphatases were shortlisted as our primary hits.

The molecules shortlisted were screened in the second screen using new siRNAs obtained from an alternative source and having sequences which were different from those used in the primary screen. All siRNAs which resulted in a Z score cut off of ± 2.0 were shortlisted as

the final hits. The final list of hits included 38 kinases and 5 phosphatases as the final hits.

An analysis of the distribution of scores obtained across the entire screen yielded a normal distribution profile with a very low ‘skewness’ value (Figure S1A). Further, a comparison of z-scores between the replicates revealed that the potency of identified targets remained

largely comparable (Figure S2). Also, both intra- and inter-plate variations for the control

wells were low with a percent standard deviation of the median value of <10%. In order to

assess the over all reproducibility of the screen we calculated the z-factor with PLK1 as

positive control. This resulted in a z-factor of 0.56 which is considered to represent an overall

high quality for the assay conditions. These collective results, therefore, confirm the

robustness, reproducibility, and sensitivity of our primary screen. (Figure S1, S2).

Population Doubling Time Experiments (PDTs).

To assess the effects of siRNA knockdown on cell cycle dynamics and population doubling

time, the following experiments were carried out with the validated siRNA hits. The siRNA

transfections were set up in duplicates and carried out in a 96-well plate format as described

above. At 48 hours post knockdown (point of maximum knockdown) the first population

doubling sample was taken and the cells were counted using trypan blue exclusion to give the

“day 0” counts. Cells were counted similarly at 72 hours and 96 hours to give the “day 1” and

“day 2” counts respectively. Here it should be emphasized that the loss of function due to

siRNA knockdown is observed up to 96 hours (Figure S1B), which is within the

recommended time frame for siRNA mediated transient knockdown experiments [1].

Population doubling times (PDTs) were calculated using the formula (log 2/log N) hr, where

N is final cell count/initial cell count (Nf/Ni) and hr is time (in hours) between initial and final cell count. The population counts were used to plot population growth curves, and also to calculate the PDTs.

Calculation of Residence Times

Using the population doubling times and the cell cycle histograms, the average residence

times (RT) of cells in different cell cycle phases were calculated as described earlier [2]. The

residence time in a particular phase G1/S/G2 (hrs) = % cells in G1/S/G1 x PDT (hrs). A table

indicating the population doubling times and respective residence times is given in Table S3.

PDTs were determined for both mock-silenced, as well as ‘hit’-silenced CH1 cells. The doubling time for mock-silenced CH1 cells was 26 hrs (SD ±3 hours), and the individual

phase-specific RTs were calculated to be G1-8.41 (SD ±1.5), S-13.48 (SD ±1), G2-4.11 (SD

±0.5). The results discussed in the text clearly indicate that each siRNA shortlisted as a ‘hit’

affected one or more cell cycle phase in terms of RT extensions. Since our interest was to

understand regulatory networks involved in G1 and S phase control, we grouped the hits

based on residence times alone. Many patterns of perturbation in RT were observed in the

resulting data. The residence times calculated are a reflection of the Z scores. A

normalization procedure was carried out to negate the plate to plate fluctuations of control

cell percentages and allow comparison between the different PDTs.

Western blot analysis

Western blot analysis was done to assess efficiency of siRNA knockdown. At appropriate

times after knockdown i.e 24 hrs, 48 hrs, 72 hrs and 96hrs, aliquots of cells were collected, centrifuged, and the cell pellets stored in liquid nitrogen. Just prior to electrophoresis, cells were lysed in lysis buffer (20mM HEPES, 10mM NaCl, 1.5nM MgCl2, 0.2mM EDTA, 0.5%

Triton X-100, 0.5mM DTT, 1mM sodium orthovanadate, 1mM NaF, and a cocktail of protease inhibitors) followed by removal of the nuclear material and other debris through centrifugation. The detergent-soluble proteins were then resolved by SDS–PAGE. Specific

proteins and phosphoproteins were detected by Western blot using appropriate antibodies.

Blots were scanned and analyzed on an Odyssey Infrared Imaging System.

Rationale for assigning cell cycle phase specific source and targets.

We identified thirty eight kinases and five phosphatases as hits from our siRNA screen

(Figure 1B, 1D). Kinases and phosphatases are generally considered as molecules involved in

signaling mechanisms. We wanted to classify the siRNA screen hits as signaling molecular

sources affecting cell cycle. Initially we could classify the hits as molecules causing

extension or no extension in cell cycle duration from the population doubling time data

(Figure 1D). Then we further classified these hits as affecting a particular phase based only

on extent of the increase in the residence time of a particular phase but not reduction in the

residence time of a particular phase to simplify further data analysis. These hits must affect the cell cycle machinery either directly or indirectly to induce such drastic changes in the

residence times of each cell cycle phases.

We defined 12 molecules as sources for causing G1 extension, 9 for S phase extension, 4 for

G2 phase extension on the basis of their residence time (Table S3). We further classified 15

molecules as both G1S phase specific sources since they affected both the phases to an equal

extent. Similarly we had 1 molecule in G1G2 and 2 in SG2 sources. By this process we

defined sources for cell cycle regulation based on our siRNA screen results and population

doubling experiments. We have included PPV scores as the ratio of confirmed predictions versus the total number of predictions for a given RNAi hit (Table S3). For this we pooled the individual phases to broadly define involvement of the siRNA ‘hits’ in the cell cycle and/or proliferation. Since it was not trivial to define the total number of predictions for a molecule available in the literature, we used Gene Ontology for the scoring purposes wherein GO terms – pertaining to biological processes – for the hits were scanned and those classes under cell cycle/proliferation (e.g. G1 regulation, mitosis

etc.) were scored as a confirmed prediction. On the other hand, the total number of GO classes for the

given molecule was considered as the total predictions.

By extensive literature curation on cell cycle we shortlisted molecules involved directly in

the regulation of cell cycle (Table S4). The target list consisted of 93 molecules grouped into

3 classes, those which are involved in regulating the G1 phase, the S phase and the G2 phase.

Many cell cycle regulators are known to play a role in G1 to S phase progression or in both

G1 and S phase regulation. These have been listed as possible G1 as well as S phase targets.

These molecules include core cell cycle regulators such as cyclins and CDKs, transcription

factors, molecules involved in DNA replication and kinases/phosphatases which are known

to play a role in cell cycle. Some of the molecules shortlisted as targets were also screened in the siRNA libraries and were shortlisted as the screen and PDT hits. These molecules may therefore appear as sources, targets and possible intermediates of the shortest paths during network analysis.

This set of 93 molecules was considered as molecular targets controlling cell cycle. Since we wished to study the phase specific regulation of cell cycle we categorized these molecules as

G1targets (51), Stargets (53) and G2targets (29) based on their involvement in regulation of a

particular phase. We combined both G1targets and Stargets molecules list to get G1Starget molecules. From the experimental and literature information we could define set of source- target relationships for phase specific cell cycle regulation.

Network analysis methods

Construction of mammalian protein-protein interaction database: We downloaded seven different mammalian protein-protein interaction (PPI) databases (i.e., reported to be present in rat/mouse/human) from the following sources, BIND [3], BioGRID [4], IntAct [5], MINT [6], HPRD [7], NetworKIN [8] and Reactome [9]. The consolidation of the different databases was achieved by merging human/mouse/rat gene symbols using the xml version of

UniprotKB downloaded from http://www.uniprot.org/downloads to yield a single PPI

database with ~12000 nodes and ~50000 edges. Since most of the experimental work has

been performed in other model systems it was imperative for us to use the collated

information to obtain a more comprehensive core network. Here it must be mentioned that

we used information from only higher organisms and have omitted interactions from Yeast,

E.coli etc. In order to incorporate all high confidence experimental/biochemical interactions,

we incorporated a filter criterion to prune out those interactions (edges) that were not

reported by at least two different experimental methods. We also removed interactions

predicted purely by in silico and high throughput methods. This yielded us a high confidence

PPI network of ~5200 nodes and ~12000 edges essentially including all known direct

experimental interactions in mammalian systems.

Shortest path analysis and ranking of cell cycle phase specific intermediates: We defined

source-target relationships involved in regulation of each of the cell cycle phases as viz.

G1sources-G1targets as G1 phase regulators, Ssources-Stargets as S phase regulators and G1Ssources-

G1Stargets as G1S phase regulators. We wished to identify key unique regulatory intermediates

(IMP nodes) for a given source-target subset for regulating a particular phase of cell cycle.

Shortest paths were calculated using Dijkstra's algorithm implemented in NetworkX. It takes

approximately 1 hour to retrieve paths in a 2GB RAM workstation. . Shortest path lengths from source to target generally varied from 3 to 7, with several such paths often being present between a given source and target combination.

Algorithm: We defined the 2 path length PPI network to be our core network (N). Given a

network N with a subset of nodes as sources (S) and targets (T) where S, T ⊂ N we had to identify the intermediate nodes and quantify their occurrences for all pairs of S and T.

1. Trace all possible shortest paths between all pairs sources S and targets T.

2. Find the intermediate nodes (I) that occur in all possible shortest paths for a given S

and T. Identify intermediate nodes I for all pairs of S and T.

3. Count the number of times each of the intermediate I occurs in shortest paths for all

pairs of S and T.

4. Compute a z-score to normalize the occurrences of an intermediate with respect to

the number of the source (S) and target (T) nodes for a given source-target subset.

nIS-T - ISS-T =

σIS-T

Where, nIS-T is the number of times the intermediate node I occurred in a given

source- target subset, is the median of number of occurrences of all

intermediates nodes in the source-target subset, σIS-T is the standard deviation of the

number of occurrences of all intermediates nodes in the source-target subset.

5. This process of identifying the intermediates and scoring was repeated for all

source-target subsets.

6. We employed a cut off criteria of intermediate score ISS-T ≥ 1.28 (p-value < 0.1) to

shortlist only significantly high scoring intermediates for our further analysis.

Comparison and short listing of intermediates with control subsets.

Since we were interested in identifying the unique phase specific IMP nodes for G1, S and

G1S phases of cell cycle we included a control criteria. Wherein molecules present in each group was compared with the distinct source-target group (as shown in the table below). The

overlapping or common molecules between these groups were filtered to retain only the

unique nodes for each phase. Here, to facilitate comparison, a similar G2source to G2target analysis was also performed and the component nodes were similarly ranked. A cut-off Z- score of >1.28 (i.e. p <0.1) was then employed for selecting the most frequently occurring nodes in all possible shortest paths for each of the cell cycle phases. From each of these individual lists, we next removed all those nodes that were also significantly present (i.e. Z- score > 1.28) in shortest paths describing the other, distinct, phases. That is, any intermediate short-listed in the G1source to G1target network was eliminated as a potential G1-specific IMP

node if this node was also over-represented in shortest paths describing either the S, or the

G2 phases. S-specific IMP nodes were delineated in a similar manner by comparing against

the high-ranking node list from the G1 and G2 phases, whereas the G1S IMP nodes were

identified by comparison against nodes from the G2 phase alone. We believe this analysis

would further yield us highly phase specific IMP nodes for subsequent network analysis.

Phase specific regulatory source-target subset Distinct control Networks

S sources-S targets

G1sources-G1targets G2 sources-G2 targets

G1sources-G1targets Ssources-S targets G2 sources-G2 targets

G1S sources -G1S targets G2sources -G2 targets

Intriguingly, with the exception of CDC25A, none of the source nodes described in Figure

2B were present in the list of extracted IMP nodes. This was in spite of the fact that shortest paths had been extracted from an undirected network thereby ensuring the absence of any inherent bias against their selection. A closer examination revealed that several of the nodes from both groups were eliminated as possible IMP nodes because their frequency of

occurrence was below the threshold value of significance. The remaining were then

eliminated in the second round of selection since these nodes were also present as frequently

occurring intermediates in the shortest paths describing the mutually exclusive cell cycle

phases. Thus these nodes were either not enriched enough, or specific enough, to be extracted

as cell cycle phase-specific IMP nodes.

Extraction of phase specific IMP node subnetworks: We wanted to extract subnetwork

regulating cell cycle phases incorporating the phase specific regulatory intermediates we

identified earlier. We considered the phase specific intermediates as the seed list and traced

all possible shortest paths of path length 2 between all pairs of nodes in the seed list. We

identified the nodes present in the shortest paths of 2 path length between all pairs of seed

nodes and merged them to create a connected subnetwork. This process was employed for all

the three subsets of phase specific intermediates yielding three such networks of 2 path

length.

Identification of cell cycle phase specific regulatory motifs: Once we identified cell cycle

phase specific subnetworks we wanted to resolve it further by manual literature curation on

all the edges present in the subnetwork. By extensive manual literature curation we could

include directions to the edges, for which the experimental evidence was present. We

compiled a flat file with human Entrez gene name of source protein as the first column, target

protein as the second column, ‘+’ for activation interaction, ‘-’ for inhibition and 0 for neutral interaction as the third. While the PMIDS for the interaction was a separate fourth column.

This network was a mixed graph of both directed and undirected edges for each of the IMP node network (Table S7). We used this network for identification of network motifs involving the phase specific IMP

nodes. We used SNAVI [10], a software tool for network analysis, for motif detection in our

literature curated IMP node subnetworks. With the phase specific regulatory intermediate

nodes as seed list and the corresponding IMP node subnetwork as background we used

SNAVI to identify the regulatory motifs for each phase separately. This yielded cell cycle phase specific motifs for G1, S and G1S phases, viz. scaffolds, bifans, feed-forward loops, feed-back loops (Figure 3D).

The number of scaffolds present in the G1S-specific IMP node network was 2-fold greater than that of G1-, and over 4-fold greater than that in the S phase-specific IMP node subnetwork (Fig. 3D). Similarly, the numbers of Feed Forward Loops (FFLs), Bifans, and

Diamond motifs were also significantly higher than the corresponding numbers present in either the G1 or the S phase sub-networks (Fig. 3D). The large number of FFLs identified in the G1S phase sub-network was particularly noteworthy. FFLs regulate both filtering of noise

and signal amplification during signal transduction, and previous studies have suggested that

they function as important transducers of signal to cell cycle responses [11]. Our findings

may thus support that FFLs serve as the crucial signal-regulatory elements driving G1 to S

phase transition. Further, the redundancy provided by over-representation of bifans and

diamond motifs could contribute towards robustness of the system against external

perturbations. Finally, our identification that the relative enrichment of signal-dependent

regulatory motifs follows the order of G1S > G1 > S is consistent with fact that the G1/S

checkpoint represents the step that is most tightly regulated by growth factor-mediated

signaling [12,13].

Network Randomization of IMP nodes sub networks: In order to validate the motifs we

identified in the phase specific IMP node networks to be functionally significant, we merged

the IMP node networks (since there are overlapping interactions) to compile a single network with mixed edges. This network was randomized over 100 times by shuffling the edge

properties. Motifs involved by the phase specific IMP nodes were scanned for in the

randomized networks using SNAVI. The figure 3E shows the results of the randomization

exercise, thus statistically validating the motifs identified.

Computing Edge densities for the motif-clusters/Modules: To calculate the edge density of

the modules, we took the IMP node network as the parent network and extracted 100 random

sub-networks with the same number of nodes as were present in our motif cluster and

calculated the edge-density for each of the random sub-networks. This was done for each of

the cell cycle phases. As is typically done, edge-density was calculated by dividing the

number of actual edges with the maximum number of possible edges. These values were then

compared with that obtained for the motif clusters and the comparison is shown in the table

below.

Edge Density=2(E)/N(N-1) Cell Cycle Phase Random Real motif cluster Z-Score subnetworks G1 0.08 0.03±0.01 4.16 G1S 0.05 0.02±0.006 4.47 S 0.38 0.10±0.05 5.66

Where E is the number of edges in the network and N is the number of nodes. It is evident from the above comparison that the edge-density of our motif clusters was at least 2.5-fold

greater than the range of values obtained for the corresponding random sub-networks. That is,

our motif clusters do exhibit at least some of the features that characterize network

communities.

Cell cycle phase specific regulatory modules and regulators: We merged the motifs we

identified for each phase to obtain a module for regulation of each phase. Since we wanted to identify the least redundant nodes in the modules, we calculated centrality parameters viz. betweenness and stress. Stress centrality defines the ability of a node to hold together communicating nodes and, therefore, is representative of the degree of its involvement in regulatory processes [14]. Betweenness, on the other hand, is a more elaborate centrality index that describes the capacity of a node to serve as a junction in the network and, thereby, regulate the network in a coherent manner [14,15]. For this we employed a selection criteria where we shortlisted only those nodes exhibiting stress and betweenness greater then 2.5 times the mean of all nodes in the corresponding phases (Fig. 4A). This we consider would identify the statistically most important nodes for further examinations in the phase specific modules.

A detailed description of the pathway model shown in Figure 6.

We wanted to characterize the growth factor/mitogen activated signaling networks that drive the cell from the preparatory G1 phase, through the G1/S transition into the DNA synthetic phase. During the G1 phase the cell does the following,

1. It senses whether the extracellular conditions are permissive for cell growth and cell

division.

2. If favourable conditions are present, it prepares to divide.

While it was initially thought that prolonged and continuous exposure to growth factors or mitogens was required to commit cells into the cycle, subsequent studies resolved that two short pulses of mitogen were sufficient to achieve this [16]. The initial pulse moved cells through the early stages of the G1 phase, while also making them responsive to the second pulse that followed several hours (~8h) later [16,17]. The second pulse then ensured progression through the later stages of G1, and consequent entry into the S phase [16]. These findings implicated the involvement of two discrete, and temporally segregated, signaling cascades in the growth factor-dependent phase of the cell cycle. The precise nature of these signaling cascades and the manner in which they exert their regulatory effects remain incompletely characterized.

Given both the functional criteria employed to segregate the source nodes for our network analysis, and subsequent verification of the stage-specific relevance of the constituent vulnerable nodes, we reasoned that the G1- and G1S-IMP node modules extracted in Figure 4 in fact represented key modules that regulated the first and second ‘wave’ of mitogen- activated signals respectively. Consequently, the discerned distribution of the vulnerability indices of these molecules between the two modules was expected to yield insights into signaling mechanisms that regulated cell cycle commitment.

In the G1 IMP node network, AKT1 and PTK2B were the two kinases that were uniquely present as high-stress high-betweenness nodes, whereas SRC and the adaptor molecule GRB2 represented the least redundant constituents of both G1 and G1S modules (Fig. 4). That is, these four molecules likely represented key constituents of the signaling cascade that governs the early G1 phase of the cycle, with AKT1 and PTK2B playing a more specific role in this process. Of these, SRC is critical for the activation of AKT1, which in turn mediates the downstream processes required for cells to enter the cycle. Although several studies implicate a central role for AKT1 in the G1 phase [18], our present findings would however restrict its involvement to the early stages of this phase. Interestingly, examination of the known properties of these vulnerable nodes supports the inference that SRC serves as a common sensor that bridges AKT1 activation with both RTKs and mitogenic GPCRs. Whereas the

SRC/GRB2 complex represents a constituent of the signalosome complex recruited by RTKs

[19], both PTK2B and the membrane-associated pool of ESR1 (another vulnerable node in the G1 IMP node sub-network) mediate SRC activation by GPCRs [20]. This may then explain earlier observations that distinct mitogenic signals activate a common signaling

cascade in the early G1 response [21].

The nodes centred on the high stress betweeness nodes seem to work together to bring about

the specific cellular outcomes. The nodes centred around the G1 high stress nodes such as

AKT and PTK2B bring about the following specific functions-

1. Driving the cell towards survival by the inhibition of apoptotic

signals/pathways.

2. Activation of Pro-survival and proliferative signals.

3. Promotion of protein synthesis.

4. Activation of pathways that increase transcription, activate, stabilize and

phosphorylate cyclin D/CDK4 complexes which can phosphorylate Rb to

initiate the cell cycle program.

The Receptor- SRC-AKT axis or early wave of signal may be required to initiate the cell cycle program by initiating the pro survival signals, promoting protein synthesis and activating survival signals. It involves the SRC dependent activation of the members of PI3K pathway, leading to the activation of AKT. Our G1 module therefore captures the critical molecules regulating key cellular processes which are needed to drive the cell through the early phase of G1.

It is evident from network of pathways shown in Figure 6 that this indeed represents an accurate synopsis of the multiplicity of biochemical events that cumulatively drive commitment of cells to the division cycle. More importantly it also reveals both the individual modes of functioning of the G1- and G1S-specific signaling cascades and the links between them that coordinate their respective activities. Thus, an examination of the relevant motifs in the merged module confirms that AKT1-dependent pathways indeed initiate cell cycle progression, but in a manner that is supported by a concomitant inhibition of pro-

apoptotic pathways, in addition to being coordinated with mechanisms regulating cell growth.

For example, activation of cell cycle regulatory pathways by AKT1 was captured through its

inhibitory effects on the CDK inhibitors p21 and p27, in addition to its positive effect on

nuclear ESR1 that is mediated through inhibition of GSK3β. The active form of nuclear

ESR1 then enhances cyclin D1 transcription (Fig. 6). Further, there is also evidence to

suggest that active AKT1 stabilizes the translated protein against proteosomal degradation

[22]. The resulting increase in cyclin D1 levels, coupled with reduced CDK inhibitory

activity, induces the formation of active cyclin D1/CDK complexes, thus marking initiation

of the cell cycle. Here, concomitant activation of the transcription factor NFkB - mediated

through AKT1-dependent phosphorylation of IKBKB - further adds by delivering a

proliferative signal. Importantly, the activation of cell cycle regulatory mechanisms by AKT1 is also well complemented by its ability to simultaneously inhibit pro-apoptotic pathways of the cell. This is achieved through the inhibition of molecules such as BAD and caspases, as well as through inhibition of activation of members of the forkhead family of transcription

factors (Fig. 6).

NFkB activation has also been implicated in regulation of cell growth by enhancing

synthesis of specific proteins [23]. Our model indicates that this effect is further augmented

through inhibition of the TSC complex by AKT1 (Fig. 6), which then activates mTOR a key

regulator of protein synthesis and cell growth [22,23,24]. Thus, Figure 6 provides an

integrated perspective on the central role played by AKT1 in driving cells through early G1.

Presumably, this predominant reliance on a single signaling intermediate to regulate the

complementary processes of cell cycle, cell growth, and promotion of cell survival, is critical

for ensuring that they all function in concert with each other. Of particular note here is that

the canonical MAP kinase pathway also supports the effector functions of AKT1 by contributing towards these three processes [24,25]. Importantly, the SRC-mediated activation

of this pathway - in a manner that is not strictly dependent upon either AKT1 or ABL1 –

allows for its recruitment in both the early and late G1 (Fig. 6). Consequently, this pathway

likely serves as at least one of the common links that facilitate transition between these two

phases.

The G1S window is regulated through cooperative interactions between the

cytoplasmic and nuclear pools of ABL1.

Figure 6 also rationalizes the key role ascribed to ABL1 during regulation of late G1 and the

subsequent G1/S transition. While such an assignment is not inconsistent with the literature,

our findings define the downstream mechanisms through which this kinase exerts its

influence. Thus, activation of ERK5 by the cytoplasmic pool of ABL1 impacts on cell cycle progression via c-fos and cyclin D1 induction [26]. This effect is further reinforced through concomitant engagement of the JNK pathway. The resulting activation of JUN then promotes the synthesis of cyclins D and A, as well as that of the transcription factor MYC [27].

Facilitation of cell cycle progression by cytoplasmic ABL1 is well supported by its ability to also suppress the pro-apoptotic functions of p53. Further, a role during actin cytoskeleton reorganization is also captured in our model through the phosphorylation of WASL and related complexes (Fig. 6).

A prominent downstream consequence of cytoplasmic ABL1 activation is the enhancement in cellular c-MYC concentrations. At one level, this is achieved through the recruitment of a Ras-dependent pathway, which influences MYC transcription [28]. In addition, ABL1-induced activation of the JAK pathway also contributes significantly to this process. Activated JAKs are potent inducers of MYC since they function both by enhancing transcription, and by stabilizing the protein through inhibition of its proteasomal degradation [28]. Thus, activation of cytoplasmic ABL1 translates into a multiplicity of mechanisms that collectively induce MYC upregulation. Here, the MAP kinase pathway also participates by phosphorylating MYC, and thereby protecting it against ubiquitinylation [28].

An important feature of our model is that it also highlights a role for the nuclear pool of

ABL1 in regulating MYC levels. A significant proportion of nuclear ABL1 is normally maintained in an inactive form through its existence in a complex with RB [28]. As cells approach the G1/S boundary however, phosphorylation of RB causes the release of ABL1. At least some fraction of this released ABL1 interacts with JUN, as a result of which ABL is activated. This process then leads to the potent activation of JUN through a feedforward regulatory mechanism wherein ABL1 directly phosphorylates JUN, and also its upstream kinase JNK [28]. Thus, signaling pathways initiated from both the cytoplasmic and nuclear pools of ABL1 converge to activate JUN and, thereby, MYC upregulation. This cooperative action between the two ABL1 subsets may well explain the long-standing question of how cellular MYC levels are temporally modulated during cell cycle progression. While expression of this protein is induced in the early G1 phase these levels are, however, further enhanced as cells prepare for S phase entry [28]. This increase is likely essential for recruiting – through transcription regulatory mechanisms - the wider spectrum of biochemical activities that are required for driving G1 to S transition. Interestingly, consistent with our proposal, the timing of this second phase of MYC enhancement coincides with that of dissociation of the RB-ABL complex [28].

Downstream of ABL1, MYC functions as a central regulator that links external signals to the cell cycle machinery. In similarity with AKT1, it suppresses pro-apoptotic pathways while also facilitating cell growth through the activation of both ribosomal RNA and protein synthesis [29,30]. Importantly, the cell cycle-regulatory effects of MYC predominantly impinge on the later stages of G1, and thereby facilitate entry of cells into the S phase. These include MYC-dependent transcription of E2Fs, positive regulation of cyclin E/CDK2

complexes, induction of Cdc25A, and the enhanced functional inactivation of CDK inhibitors

[30]. Indeed the obligatory requirement for MYC during S phase entry is underscored by

results demonstrating that either a reduction in its levels, or inhibition of its activity, causes a

block in G1/S but not G0/G1 transition [30].

Thus the above analysis reiterates that our G1 and G1S IMP node modules indeed capture the core set of signaling processes that guide cells during the commitment phase of the cell cycle. Further, it also provides a mechanism-based rationalization of our earlier categorization - on the basis of graph theoretical considerations – of some of the nodes as displaying a high degree of vulnerability. Finally, this rationalization also extends to our assignment of key regulatory roles to AKT1 and ABL1 in the early and late G1 stages respectively.

Coordination of signal modulation with cell cycle progression.

A resolution to the issue of how ABL1 replaces AKT1 as the dominant signaling node in the

G1S phase was also suggested from an analysis of the pathways captured by our model (Fig.

6). As illustrated, this process hinges upon the release of nuclear ABL1 from its complex with RB. AKTI-dependent initiation of cells into the cycle culminates with the phosphorylation of RB, which is normally evident by 8h after stimulation of cells with a mitogenic signal [17]. This leads to the release of bound ABL1. As noted earlier, activation of this subset synergizes with cytoplasmic ABL signaling to ensure enhancement in MYC levels. In addition, some fraction of this nuclear ABL1 is also likely to partition into the cytoplasmic compartment of the cell. This could be achieved either through its direct export in the non-phosphorylated form, or, following acetylation by the histone acetyltransferase EP300 [31]. The resulting increase in the pool of cytoplasmic ABL1 would then further intensify signals generated from this subset of the kinase.

Thus, our model indicates a marked amplification of ABL1-dependent signaling as cells approach the G1/S boundary. This occurs through both enrichment of the cytoplasmic pool, as well as through activation of the fraction that remains in the nucleus. Importantly, this amplification would be further reinforced by the fact that signals from both subsets synergize at the level of activating downstream effector pathways such as MYC regulation, and the synthesis of cyclins. These cumulative effects may then explain how ABL1 acquires the characteristics of a vulnerable node in the later stages of G1. It is pertinent to recall here that although AKT1 is present in both the G1 and G1S modules, its properties of high-stress and high-betweenness are restricted to the G1 module (Fig. 4). That is, in the late G1 phase,

ABL1 replaces AKT1 as the obligatory signaling intermediate. Significantly, this switch in signaling axes occurs in a seamless manner because, in addition to recruiting the additional pathways required for S phase entry, ABL1 also supplants the functions of AKT1 in terms of suppressing apoptotic pathways and regulating cell growth. This would then explain the increased functional redundancy of AKT1 at this stage.

Our proposal that signals from nuclear ABL1 contribute towards promoting cell cycle progression is apparently inconsistent with earlier findings that this subset primarily functions to activate the cellular apoptotic program in response to conditions such as genotoxic stress

[32]. This latter activity, however, is strictly dependent upon its interactions with p53 and p73. As earlier discussed, mitogenic signals cause suppression of pro-apototic pathways, which include inhibition of activities of both p53 and p73. This is initiated in an AKT1- dependent manner in early G1, and then subsequently sustained by signals generated from cytoplasmic ABL1. We thus propose that activation of nuclear ABL1 under conditions where pro-apoptotic pathways are inhibited biases its role towards integrating with the cell cycle stimulatory functions of its cytoplasmic counterpart.

REFERENCES.

1. Echeverri CJ, Perrimon N (2006) High-throughput RNAi screening in cultured cells: a user's guide. Nat Rev Genet 7: 373-384. 2. Neufeld TP, de la Cruz AF, Johnston LA, Edgar BA (1998) Coordination of growth and cell division in the Drosophila wing. Cell 93: 1183-1193. 3. Bader GD, Donaldson I, Wolting C, Ouellette BF, Pawson T, et al. (2001) BIND--The Biomolecular Interaction Network Database. Nucleic Acids Res 29: 242-245. 4. Stark C, Breitkreutz BJ, Reguly T, Boucher L, Breitkreutz A, et al. (2006) BioGRID: a general repository for interaction datasets. Nucleic Acids Res 34: D535-539. 5. Kerrien S, Alam-Faruque Y, Aranda B, Bancarz I, Bridge A, et al. (2007) IntAct--open source resource for molecular interaction data. Nucleic Acids Res 35: D561-565. 6. Chatr-aryamontri A, Ceol A, Palazzi LM, Nardelli G, Schneider MV, et al. (2007) MINT: the Molecular INTeraction database. Nucleic Acids Res 35: D572-574. 7. Keshava Prasad TS, Goel R, Kandasamy K, Keerthikumar S, Kumar S, et al. (2009) Human Protein Reference Database--2009 update. Nucleic Acids Res 37: D767-772. 8. Linding R, Jensen LJ, Pasculescu A, Olhovsky M, Colwill K, et al. (2008) NetworKIN: a resource for exploring cellular phosphorylation networks. Nucleic Acids Res 36: D695-699. 9. Matthews L, Gopinath G, Gillespie M, Caudy M, Croft D, et al. (2009) Reactome knowledgebase of human biological pathways and processes. Nucleic Acids Res 37: D619-622. 10. Ma'ayan A, Jenkins SL, Webb RL, Berger SI, Purushothaman SP, et al. (2009) SNAVI: Desktop application for analysis and visualization of large-scale signaling networks. BMC Syst Biol 3: 10. 11. Csikasz-Nagy A, Kapuy O, Toth A, Pal C, Jensen LJ, et al. (2009) Cell cycle regulation by feed-forward loops coupling transcription and phosphorylation. Mol Syst Biol 5: 236. 12. Sears RC, Nevins JR (2002) Signaling networks that link cell proliferation and cell fate. J Biol Chem 277: 11617-11620. 13. Zetterberg A, Larsson O, Wiman KG (1995) What is the restriction point? Curr Opin Cell Biol 7: 835-842. 14. Manimaran P, Hegde SR, Mande SC (2009) Prediction of conditional gene essentiality through graph theoretical analysis of genome-wide functional linkages. Mol Biosyst. 15. Newman MEJ (2005) A measure of betweenness centrality based on random walks. Social Networks 27: 39-54. 16. Jones SM, Kazlauskas A (2001) Growth-factor-dependent mitogenesis requires two distinct phases of signalling. Nat Cell Biol 3: 165-172. 17. Ezhevsky SA, Ho A, Becker-Hapak M, Davis PK, Dowdy SF (2001) Differential regulation of retinoblastoma tumor suppressor protein by G(1) cyclin-dependent kinase complexes in vivo. Mol Cell Biol 21: 4773-4784. 18. Massague J (2004) G1 cell-cycle control and cancer. Nature 432: 298-306. 19. Luttrell DK, Luttrell LM (2004) Not so strange bedfellows: G-protein-coupled receptors and Src family kinases. Oncogene 23: 7969-7978. 20. Litvak V, Tian D, Shaul YD, Lev S (2000) Targeting of PYK2 to focal adhesions as a cellular mechanism for convergence between integrins and G protein-coupled receptor signaling cascades. J Biol Chem 275: 32736-32746. 21. Jones SM, Kazlauskas A (2000) Connecting signaling and cell cycle progression in growth factor-stimulated cells. Oncogene 19: 5558-5567. 22. Gera JF, Mellinghoff IK, Shi Y, Rettig MB, Tran C, et al. (2004) AKT activity determines sensitivity to mammalian target of rapamycin (mTOR) inhibitors by regulating cyclin D1 and c-myc expression. J Biol Chem 279: 2737-2746. 23. Dan HC, Cooper MJ, Cogswell PC, Duncan JA, Ting JP, et al. (2008) Akt-dependent regulation of NF-{kappa}B is controlled by mTOR and Raptor in association with IKK. Genes Dev 22: 1490-1500. 24. Zhou BP, Liao Y, Xia W, Spohn B, Lee MH, et al. (2001) Cytoplasmic localization of p21Cip1/WAF1 by Akt-induced phosphorylation in HER-2/neu-overexpressing cells. Nat Cell Biol 3: 245-252. 25. Vincent AM, Feldman EL (2002) Control of cell survival by IGF signaling pathways. Growth Horm IGF Res 12: 193-197. 26. Brown JR, Nigh E, Lee RJ, Ye H, Thompson MA, et al. (1998) Fos family members induce cell cycle entry by activating cyclin D1. Mol Cell Biol 18: 5609-5619. 27. Obaya AJ, Mateyak MK, Sedivy JM (1999) Mysterious liaisons: the relationship between c-Myc and the cell cycle. Oncogene 18: 2934-2941. 28. Sirvent A, Benistant C, Roche S (2008) Cytoplasmic signalling by the c-Abl tyrosine kinase in normal and cancer cells. Biol Cell 100: 617-631. 29. Mateyak MK, Obaya AJ, Sedivy JM (1999) c-Myc regulates cyclin D-Cdk4 and -Cdk6 activity but affects cell cycle progression at multiple independent points. Mol Cell Biol 19: 4672-4683. 30. Amati B, Alevizopoulos K, Vlach J (1998) Myc and the cell cycle. Front Biosci 3: d250- 268. 31. di Bari MG, Ciuffini L, Mingardi M, Testi R, Soddu S, et al. (2006) c-Abl acetylation by histone acetyltransferases regulates its nuclear-cytoplasmic localization. EMBO Rep 7: 727-733. 32. Dierov J, Dierova R, Carroll M (2004) BCR/ABL translocates to the nucleus and disrupts an ATR-dependent intra-S phase checkpoint. Cancer Cell 5: 275-285.